Metatron Intelligence
Today: the five AI headlines of the day → and the AI Wiki
What a domain is How to read this page

One of twenty-four connected subject areas

Domains are how claims relate to one another inside the Atlas. Desks are how readers find them. A claim can sit in more than one domain without being counted twice.

Coverage here is not importance

A domain with more claims has been worked on more, which is a fact about this publication rather than a fact about the world.

Everything returns to the record

Each entry links to its canonical claim record, with its source, its non-claim and its update condition intact.

New to this publication?

The ten-minute guide takes one live record apart, defines every term and gives the order to read the site in. Start here →

Business · D03

Intelligence, Research & Decision Products

500 connected claims · 306 source-domain links · Priority P0

Search and filter this domain →
  1. MWITA-CR-2026-0026Lokale Sprache ist Infrastruktur: SEA-LION, BHASHINI, Masakhane und Latam-GPT adressieren Daten-, Benchmark-, Code-Switching-, Rechte- und Repräsentationslücken, die eine nachträgliche Übersetzung nicht zuverlässig löst.Evidence B · P0
  2. MWITA-CR-2026-0028Der Begriff Global South ist für Produkt- und Marktentscheidungen zu grob, weil kapital- und infrastrukturdichte Hubs, mobile-money-/agentengetriebene Märkte, geschlossene Plattformökosysteme und Instant-Payment-Leitmärkte fundamental unterschiedliche Systemzustände darstellen.Evidence B · P0
  3. MWITA-KI-2026-012RBI reported its Digital Payments Index at 465.33 for September 2024, versus 445.50 for March 2024, on a March 2018 base of 100.Evidence A · P0
  4. MWITA-KI-2026-013RBI reported its Digital Payments Index at 493.22 for March 2025, versus 465.33 for September 2024, on a March 2018 base of 100.Evidence A · P0
  5. MWITA-KI-2026-014RBI reported its Digital Payments Index at 516.76 for September 2025, versus 493.22 for March 2025, on a March 2018 base of 100.Evidence A · P0
  6. MWITA-KI-2026-021Government of India releases reported IndiaAI compute capacity exceeding 45,000 GPUs in 2026.Evidence B · P1
  7. MWITA-KI-2026-023MoSPI changed the PLFS methodology from January 2025 and began monthly all-India current-weekly-status estimates plus quarterly State/UT estimates.Evidence A · P0
  8. MWITA-DM-2026-001A strictly proper scoring rule maximizes a forecaster's expected score when the reported distribution equals the forecaster's true belief; Brier's quadratic score is a foundational binary-event implementation.Evidence A · P0
  9. MWITA-DM-2026-002Probabilistic forecast quality must consider calibration and sharpness jointly; a forecaster can be calibrated while remaining uninformative by repeatedly issuing the base rate.Evidence A · P0
  10. MWITA-DM-2026-003The US IARPA ACE program treated probabilistic elicitation, weighting and aggregation as measurable research objects and found that structured crowd-forecasting methods could be compared by accuracy.Evidence B · P2
  11. MWITA-DM-2026-013NIST AI RMF treats trustworthy AI as a combination of validity/reliability, safety, security/resilience, accountability/transparency, explainability/interpretability, privacy and fairness; the Measure function calls for documented, objective, repeatable or scalable TEVV.Evidence A · P0
  12. MWITA-DM-2026-014NIST's Generative AI Profile centers governance, content provenance, pre-deployment testing and incident disclosure and treats risks across model, application and ecosystem levels.Evidence A · P0
  13. MWITA-DM-2026-015HELM demonstrates a reproducible multi-scenario, multi-metric evaluation architecture and explicitly records missing scenarios/metrics; it does not claim that one model dominates all desiderata.Evidence B · P1
  14. MWITA-DM-2026-017NIST's 2026 agentic-AI probe research evaluates evidence attribution using separable dimensions of faithfulness, completeness and sufficiency and produces structured audit trails.Evidence C · P1
  15. MWITA-DM-2026-018Strategic foresight explores multiple plausible futures rather than predicting one future, and its operational value comes from stress-testing current policy and identifying robust or contingent actions.Evidence A · P1
  16. MWITA-DM-2026-019UK government guidance defines horizon scanning as systematic collection of trends and weak signals and recommends looking beyond usual timescales, sources and culture, followed by organized storage, analysis and write-up.Evidence B · P2
  17. MWITA-DM-2026-020Metatron will learn more reliably when forecasts, evidence, interventions, decisions and outcomes share stable identifiers and append-only histories than when reports are stored as disconnected documents.Evidence D · P0
  18. MWITA-AW-2026-001In Bank of Korea authors' out-of-sample evaluation from 2016 onward, PRISM-Now outperformed comparison ensemble methods based on ex-post information, simple means, or medians.Evidence B · P1
  19. MWITA-AW-2026-002PRISM-Now's evaluation found conventional core-macro models stronger for current-quarter nowcasts when information was nearly complete, while big-data and nonlinear ML models were stronger one quarter ahead when realized information was absent.Evidence B · P1
  20. MWITA-AW-2026-003RBI reported an AI/ML system in FY2024-25 that regularly tracks and analyses corporate sentiment from online and print media to supplement its enterprise surveys.Evidence B · P1
  21. MWITA-AW-2026-004RBI's FY2024-25 report classifies LLM use for short-term forecasting as experimental and reports prototypes for CPI forecasting plus AI/ML models for onion- and wheat-CPI using unstructured and traditional data.Evidence B · P1
  22. MWITA-AW-2026-005Banco Central do Brasil reported that its SupTech programme created ten NLP-based tools in use across supervisory activities and also established AI-governance rules for development, approval, use, and maintenance.Evidence B · P1
  23. MWITA-AW-2026-006Bank of Korea's 2025 BOK-LOOK working paper presents a semi-structural model designed to produce internally consistent forecasts and counterfactual scenarios, explicitly including household-debt and housing-price dynamics.Evidence B · P1
  24. MWITA-AW-2026-019The FREE-AI report characterizes agentic systems as a potential shift from task automation to decision automation and gives an SME lending-agent scenario, but presents it as prospective rather than evidence of current deployment.Evidence B · P1
  25. MWITA-LC-2026-005Colombia approved CONPES 4144 in February 2025 as its national artificial-intelligence policy, with an action plan spanning governance, data and infrastructure, research and innovation, capabilities and risk mitigation.Evidence A · P1
  26. MWITA-RD-2026-003Singapore's IMDA launched version 1.0 of its Model AI Governance Framework for Agentic AI on 22 January 2026, recommending risk-bounded autonomy, meaningful human approval points, lifecycle technical controls and end-user transparency.Evidence A · P1
  27. MWITA-WD-2026-013Canada launched its first AI Strategy for the federal public service in March 2025; it is a government-operating roadmap rather than a law governing all Canadian private-sector AI.Evidence A · P1
  28. MWITA-WD-2026-017ONS estimated in June 2026 that about 35% of UK businesses with 10 or more employees reported using at least one AI technology, but adoption depth rose only from about 1.4 to 1.6 technologies per adopting business since 2023 and had not produced widespread workforce-headcount change.Evidence A · P1
  29. MWITA-WD-2026-019Ofsted's June 2025 study of 21 early-adopter schools, colleges and trusts found workload-oriented AI uses and governance practices, but concluded that robust evidence of educational impact remained limited, explorative and short-term.Evidence A · P1
  30. MWITA-BM-2026-001Salesforce reported Agentforce annual recurring revenue of $800 million for fiscal 2026, up 169% year over year; this is a company-defined product-adoption financial metric reported in an SEC filing.Evidence A · P1
  31. MWITA-BM-2026-002Salesforce reported 2.4 billion Agentic Work Units delivered as of fiscal Q4 2026, using a company-defined unit for agentic work consumption.Evidence A · P1
  32. MWITA-BM-2026-004Intercom lists Fin AI Agent from $0.99 per outcome; an outcome can be a confirmed or assumed resolution or a completed configured workflow, and billing occurs at most once per conversation.Evidence B · P1
  33. MWITA-BM-2026-005Salesforce lists Agentforce Flex Credits at $500 per 100,000 credits, with a standard agent action consuming 20 credits, and also offers $2-per-conversation pricing.Evidence B · P1
  34. MWITA-BM-2026-006Zendesk defines automated resolutions as the billing unit for AI-agent usage and describes payment as occurring when a request is resolved without escalation to a human agent.Evidence B · P1
  35. MWITA-BM-2026-010Palantir reported $4.475 billion total revenue and 954 customers in 2025; commercial revenue was $2.073 billion, up 60% year over year, while AIP is bundled with its existing platforms.Evidence A · P1
  36. MWITA-BM-2026-011Snowflake reported $4.47 billion fiscal-2026 product revenue, $9.77 billion remaining performance obligations and 13,328 customers at 31 January 2026.Evidence A · P1
  37. MWITA-BM-2026-013Appier reported JPY43.7 billion fiscal-2025 revenue, up 28%, JPY3.0 billion operating profit and 13% year-over-year growth in its customer base.Evidence A · P1
  38. MWITA-BM-2026-017Salesforce stated that its internal Agentforce deployment resolved 500,000 support cases and handled 84% without human intervention.Evidence C · P1
  39. MWITA-MB-2026-001Meridian explicitly states that out-of-sample prediction metrics are only structural checks and cannot directly validate causal inference; correctly aligned experiments are needed to assess causal marketing effects.Evidence A · P1
  40. MWITA-MB-2026-002Meridian recommends using incrementality experiments to calibrate channel ROI priors, while warning that experiment and MMM must target the same causal quantity and time/channel scope.Evidence A · P1
  41. MWITA-MB-2026-003GeoLift implements power analysis, market selection and synthetic-control inference for aggregated geo experiments; its example outputs are simulated demonstrations, not evidence that any real campaign generated the displayed lift.Evidence A · P1
  42. MWITA-MB-2026-004GeoLift's multi-cell workflow shows that two cells can each have significant lift while their wide comparative uncertainty prevents declaring one strategy the winner.Evidence A · P1
  43. MWITA-MB-2026-005METR randomized 246 tasks from 16 experienced open-source developers on familiar repositories and estimated that access to early-2025 AI tools increased completion time by 19%.Evidence A · P1
  44. MWITA-MB-2026-006METR's 2026 design update reports increasing refusal to work without AI, which can bias estimates when participants willing to enter a no-AI condition differ from the target population.Evidence A · P1
  45. MWITA-MB-2026-007METR defines task-completion time horizon as the human-duration point at which an agent is predicted to succeed at a specified reliability under a particular task suite and scaffold.Evidence A · P1
  46. MWITA-MB-2026-008τ-bench introduced pass^k to measure the probability that an agent succeeds consistently across repeated tool-user interactions, exposing reliability that single-attempt success can hide.Evidence A · P1
  47. MWITA-MB-2026-009τ²-bench models both agent and user as actors who can change shared state, separating reasoning errors from communication and coordination errors in telecom scenarios.Evidence B · P1
  48. MWITA-MB-2026-010An ICML 2025 study formalized detection of benchmark leakage and showed why overlap between evaluation data and training/fine-tuning data undermines performance interpretation.Evidence A · P1
  49. MWITA-MB-2026-011The 2025 international joint testing exercise coordinated AI safety institutes to refine common agentic-evaluation practices; its stated goal was evaluation science and shared methods, not certifying models as safe.Evidence A · P1
  50. MWITA-MB-2026-012Anthropic found harmful choices across models in deliberately constructed corporate simulations but explicitly reported no evidence of this behavior in real deployments.Evidence B · P1
  51. MWITA-MB-2026-013SHADE-Arena evaluates whether monitors can detect covert sabotage in complex agent traces, making monitoring recall, false positives and review burden measurable rather than assumed.Evidence B · P1
  52. MWITA-MB-2026-014OpenAI's Operator system card describes layered safeguards for computer use, including confirmation and risk controls, while also reporting model mistakes and prompt-injection risks.Evidence B · P1
  53. MWITA-MB-2026-015A 2025 QJE study of staggered rollout to 5,172 support agents estimated 15% more resolved issues per hour, with larger gains for less-experienced agents and small quality declines among some highly skilled agents.Evidence A · P1
  54. MWITA-MB-2026-016A 2026 Alibaba field experiment reported improved customer ratings and dissatisfaction measures with generative AI while finding no significant effect on objective customer retrial rates.Evidence B · P1
  55. MWITA-MB-2026-017A World Bank randomized trial of a six-week after-school program combining Copilot with teacher guidance estimated a 0.31-standard-deviation gain on a combined assessment among Nigerian secondary students.Evidence A · P1
  56. MWITA-MB-2026-018A large preregistered Turkish high-school RCT found that standard GPT-4 access improved assisted practice performance but reduced students' unassisted exam performance; a teacher-designed guarded tutor mitigated the harm.Evidence A · P1
  57. MWITA-MB-2026-019A 2025 randomized study of 194 college students found greater immediate learning in less time with a research-designed AI tutor than with a matched active-learning class.Evidence A · P1
  58. MWITA-MB-2026-020A 2025 randomized surgical-simulation trial found expert instructors augmented with AI performance data produced better skill acquisition and transfer than an intelligent tutor condition.Evidence A · P1
  59. MWITA-MB-2026-021In a 16-facility cluster-randomized trial with 9,691 patients, generative-AI decision support did not significantly reduce the prespecified 14-day treatment-failure composite; rare patient outcomes required far larger samples.Evidence A · P1
  60. MWITA-MB-2026-022FDA's 2025 request for comment states that retrospective tests and static benchmarks provide a baseline but are not designed to predict dynamic real-world behavior, motivating ongoing monitoring for drift and performance variation.Evidence A · P1
  61. MWITA-MB-2026-023CONSORT 2025 and SPIRIT 2025 specify minimum reporting for randomized trials; AI extensions require algorithm version, inputs, outputs, error handling and human-AI interaction to be described.Evidence A · P1
  62. MWITA-AI-2026-002OpenAI reported that offline evaluations and A/B tests looked favorable before the sycophancy rollback, while the specific harmful behavior was not represented in deployment gates.Evidence A · P1
  63. MWITA-AI-2026-017SHADE-Arena evaluates whether monitors detect deliberately hidden sabotage in controlled tasks, but benchmark detection does not establish that deployed monitors cover real workflows or novel attacks.Evidence B · P1
  64. MWITA-AI-2026-018FCA consumer pilots found that chatbot users could report willingness or confidence even when average task performance, including savings-account selection or challenges to incorrect AI credit decisions, did not improve commensurately.Evidence A · P1
  65. MWITA-WO-2026-028Anthropic's January 2026 Economic Index introduced success-weighted effective task coverage and showed that occupation rankings change when task frequency, duration and model success are incorporated.Evidence B · P1
  66. MWITA-FB-2026-018Kao reported that its expanded Kirei Skin AI used 25 models to estimate 77 skin attributes from facial images, processed each suitable image in under ten minutes, and was analyzing about 730 images per month in its research operations in 2025.Evidence B · P1
  67. MWITA-PW-2026-001China reported real-estate development investment of CNY8.2788 trillion in 2025, 17.2% below 2024.Evidence A · P1
  68. MWITA-PW-2026-002Chinese real-estate developers newly started 587.70 million square metres of floor space in 2025, 20.4% less than in 2024.Evidence A · P1
  69. MWITA-PW-2026-003Chinese developers reported 603.48 million square metres of completed floor space in 2025, down 18.1% year on year.Evidence A · P1
  70. MWITA-PW-2026-004China reported 881.01 million square metres of newly built commercial buildings sold by contract in 2025, 8.7% less than in 2024.Evidence A · P1
  71. MWITA-PW-2026-005At end-2025, China reported 766.32 million square metres of completed commercial-building floor space available for sale or lease, 1.6% more than a year earlier.Evidence A · P1
  72. MWITA-PW-2026-006URA reported 11,482 uncompleted private residential units launched for sale in 2025, up from 6,647 in 2024.Evidence A · P1
  73. MWITA-PW-2026-007URA reported developers sold 10,815 private residential units excluding executive condominiums in 2025, up from 6,469 in 2024.Evidence A · P1
  74. MWITA-PW-2026-008URA reported 14,622 private residential resale transactions in 2025 versus 14,053 in 2024.Evidence A · P1
  75. MWITA-PW-2026-009URA reported 7,996 private residential and executive-condominium units completed in 2025.Evidence A · P1
  76. MWITA-PW-2026-010URA reported the vacancy rate of completed private residential units excluding executive condominiums fell from 6.9% in Q3 to 6.0% at end-Q4 2025.Evidence A · P1
  77. MWITA-PW-2026-011URA reported office-space prices fell 2.1% over 2025 while office rents rose 0.3%.Evidence A · P1
  78. MWITA-PW-2026-012Saudi Arabia's GASTAT reported the national real-estate price index declined 0.7% in Q4 2025 from Q4 2024, with residential prices down 2.2%.Evidence A · P1
  79. MWITA-PW-2026-013GASTAT states that its updated quarterly real-estate price index integrates transaction data from all administrative regions with multiple sources and geospatial AI models.Evidence A · P1
  80. MWITA-PW-2026-014RBI reported the All-India House Price Index rose 3.1% year on year in Q4 2024-25 and 0.9% sequentially.Evidence A · P1
  81. MWITA-PW-2026-015RBI reported Q4 2024-25 annual house-price change ranging from an 8.8% increase in Kolkata to a 2.3% contraction in Kochi.Evidence A · P1
  82. MWITA-PW-2026-016Statistics South Africa reported the nominal value of private-sector building plans passed by larger municipalities fell 2.6% in 2025.Evidence A · P1
  83. MWITA-PW-2026-017Statistics South Africa reported the nominal value of buildings completed rose 2.6% in 2025, driven by higher residential completion value while non-residential completion value fell.Evidence A · P1
  84. MWITA-IFD-2026-001The African Development Bank Group reported more than USD11 billion of financing approvals in 2025.Evidence A · P1
  85. MWITA-IFD-2026-002The African Development Bank Group reported USD7.1 billion in 2025 disbursements, its highest annual level in five years.Evidence A · P1
  86. MWITA-IFD-2026-003AfDB reported that its 2025 financing approvals covered 200 projects across 50 African countries.Evidence A · P1
  87. MWITA-IFD-2026-004AfDB's 2025 annual-results communication reports that 1.5 million people gained access to electricity through bank-supported operations.Evidence B · P1
  88. MWITA-IFD-2026-005AfDB reports that 6.9 million farmers adopted climate-resilient technologies in its 2025 development results.Evidence B · P1
  89. MWITA-IFD-2026-006The New Development Bank reports approving 19 projects for USD3.171 billion in financing during 2025.Evidence A · P1
  90. MWITA-IFD-2026-007At the end of 2025, NDB reported 115 projects with USD35.593 billion in approved financing.Evidence A · P1
  91. MWITA-IFD-2026-008NDB estimates that projects approved in 2025 are expected to install 2,200 MW of renewable and clean-energy generation capacity.Evidence B · P1
  92. MWITA-IFD-2026-009NDB estimates projects approved in 2025 could avoid 17.6 million tonnes of carbon dioxide emissions annually.Evidence B · P1
  93. MWITA-IFD-2026-010ADB reports USD1.03 billion in sovereign and nonsovereign commitments to Nepal during 2025, including USD86 million in Asian Development Fund grants.Evidence A · P1
  94. MWITA-IFD-2026-011ADB reports cumulative public- and private-sector loan and grant disbursements to Nepal of USD6.09 billion as of the 2025 country profile.Evidence A · P1
  95. MWITA-IFD-2026-012AIIB's 2025 business plan defines project implementation readiness using the number of months between project approval and first cash disbursement.Evidence A · P1
  96. MWITA-IFD-2026-013AIIB's 2025 business-plan methodology excludes policy-based financing from its annual disbursement ratio and implementation-readiness measures because such operations would distort them.Evidence A · P1
  97. MWITA-IFD-2026-014The Green Climate Fund reported Board approval of 50 projects totaling USD3.26 billion in 2025.Evidence A · P1
  98. MWITA-IFD-2026-015GCF's project page for Kyrgyz Republic project SAP002 reports USD8,576,108 in GCF grant financing as 100% disbursed.Evidence A · P1
  99. MWITA-IFD-2026-016GCF lists 802,000 people as the anticipated number with increased resilience for SAP002 in the Kyrgyz Republic.Evidence B · P1
  100. MWITA-IFD-2026-017GCF's Mozambique SAP011 project page reports USD1,910,730 as 100% disbursed.Evidence A · P1
  101. MWITA-IFD-2026-018The Global Fund Results Report 2025 reports 25.6 million people on antiretroviral therapy in 2024 in countries and regions where it invests.Evidence B · P1
  102. MWITA-IFD-2026-019The Global Fund reports 70 million lives saved since 2002 across countries where it invests, using disease-burden and impact estimates from WHO, UNAIDS and modeling partners.Evidence B · P1
  103. MWITA-IFD-2026-020Gavi reports that zero-dose children in 57 supported lower-income countries declined to 10.2 million in 2024, 5% below 2023.Evidence B · P1
  104. MWITA-IFD-2026-021Gavi reports DTP3 immunization coverage of 82% in supported countries in 2024, up from 78% in 2021 and near the 83% pre-pandemic level in 2019.Evidence B · P1
  105. MWITA-IFD-2026-022OECD preliminary data put 2025 official development assistance from DAC members and associates at USD174.3 billion, down 23.1% in real terms from 2024.Evidence A · P1
  106. MWITA-PB2B-2026-001India's Government e Marketplace reported gross merchandise value above INR5 lakh crore for fiscal year 2025-26.Evidence A · P1
  107. MWITA-PB2B-2026-002GeM reported cumulative gross merchandise value of INR18.4 lakh crore since inception by 6 April 2026.Evidence A · P1
  108. MWITA-PB2B-2026-003The Department of Commerce reported more than 24 lakh profile-completed sellers and service providers on GeM as of 30 November 2025.Evidence A · P1
  109. MWITA-PB2B-2026-004GeM reported more than 1.67 lakh buyer organizations on the platform as of 30 November 2025.Evidence A · P1
  110. MWITA-PB2B-2026-005India reported that micro and small enterprises accounted for 44.8% of cumulative GeM order value and had received orders worth more than INR7.35 lakh crore by 30 November 2025.Evidence A · P1
  111. MWITA-PB2B-2026-006A Government of India release reports that an IIT Delhi study covering FY2023-24 through FY2025-26 estimated monetized price and process benefits of INR86,571.69 crore on GeM.Evidence B · P1
  112. MWITA-PB2B-2026-007The same government release says the IIT Delhi study found GeM prices lower for 74.3% of 101 items compared with other e-marketplaces, with a spend-weighted saving of 13.6%.Evidence B · P1
  113. MWITA-PB2B-2026-008ChileCompra reports 1,868,361 purchase orders issued by 1,165 public entities through Mercado Público in 2025.Evidence A · P1
  114. MWITA-PB2B-2026-009ChileCompra reports 100,407 suppliers that submitted offers, quoted or received a purchase order in 2025; 83% were smaller enterprises.Evidence A · P1
  115. MWITA-PB2B-2026-010ChileCompra explicitly states that the purchase orders used to calculate transacted amounts are commitments and not payments by public bodies.Evidence A · P1
  116. MWITA-PB2B-2026-011ChileCompra reported preliminary 2025 public-procurement savings of USD381,049,227.Evidence B · P1
  117. MWITA-PB2B-2026-012ChileCompra reports that direct awards fell from 20% to 12.5% of transacted value during 2025.Evidence A · P1
  118. MWITA-PB2B-2026-013ChileCompra's consolidated methodology estimates USD1.191 billion in savings across public tenders, framework agreements and coordinated purchases during 2023-2025.Evidence B · P1
  119. MWITA-PB2B-2026-014ChileCompra reports that smaller enterprises transacted USD7.841 billion in 2025, equal to 35.6% of national public-procurement value, and received about six in ten purchase orders.Evidence A · P1
  120. MWITA-PB2B-2026-015As updated 14-15 August 2026, South Africa's National Treasury transparency portal displayed 158,295 tenders published and 66,872 awards, with coverage beginning 15 August 2025 for several dashboard views.Evidence A · P1
  121. MWITA-PB2B-2026-016South Africa's National Treasury transparency portal displayed 1,461,643 suppliers in August 2026.Evidence A · P1
  122. MWITA-PB2B-2026-017Korea's Public Procurement Service describes a public-procurement market of about KRW225 trillion, connecting roughly 620,000 suppliers and 75,000 end-user institutions.Evidence A · P1
  123. MWITA-PB2B-2026-018Korea's PPS states that a supplier registers once in KONEPS to participate in public bidding opportunities and that the system covers the process from tender information through contracting and payment.Evidence A · P1
  124. MWITA-PB2B-2026-019Saudi Arabia's national platform reports more than 885,000 contracts and awards and SAR1 trillion in payment-order value through Etimad.Evidence A · P1
  125. MWITA-PB2B-2026-020The UAE Ministry of Finance states that catalogue ordering on its Digital Procurement Platform reduced procurement process time from 60 days to 6 minutes.Evidence B · P1
  126. MWITA-PB2B-2026-021Australia's Department of Finance reports 86,926 contracts published on AusTender for 2024-25 with a combined reported value of AUD104.9 billion.Evidence A · P1
  127. MWITA-PB2B-2026-022For 2024-25, Australia estimated 45,556 SME contracts, 52.4% of contract volume, while SMEs received 20.7% of contract value below AUD1 billion, totaling AUD14.7 billion.Evidence A · P1
  128. MWITA-PB2B-2026-023UK procurement guidance states that from 24 February 2025 public procurement contracts imply a 30-day term for valid, undisputed invoices, extending through qualifying subcontracts.Evidence A · P1
  129. MWITA-DA-2026-006India's 2025 AT Rules direct creation of a national safety-incident database for AT dysfunctions causing injury or accidents and a multimodal grievance system with a 30-day resolution timeline.Evidence A · P1
  130. MWITA-DA-2026-015Australia's Digital Inclusion Standard is marked mandatory and applicable from 1 January 2025 to covered investments and existing services under the Digital Experience Policy.Evidence A · P1
  131. MWITA-DA-2026-019NSW's Digital Inclusion Strategy lists embedding digital-accessibility requirements in government procurement as an initiative.Evidence B · P1
  132. MWITA-DA-2026-034Google announced in May 2025 that TalkBack users could ask Gemini questions about images and whole-screen content in addition to receiving generated descriptions.Evidence B · P1
  133. MWITA-SI-2026-021NPCI reports 731 banks live on UPI and 22.716 billion transactions worth INR28.921 trillion in June 2026.Evidence A · P1
  134. MWITA-SI-2026-022NPCI's August 2025 remitter statistics show approval rates among major UPI banks ranging from below 86% to above 95%, with debit-reversal success also varying substantially.Evidence A · P1
  135. MWITA-SI-2026-023Banco Central do Brasil reports that Pix processed nearly 80 billion transactions worth more than BRL35 trillion in 2025.Evidence A · P1
  136. MWITA-SI-2026-024Banco Central do Brasil reports 148 million individuals and 12.8 million companies had sent or received at least one Pix transaction by the end of 2025.Evidence A · P1
  137. MWITA-SI-2026-025Banco Central do Brasil's Pix statistics listed 889 participants as of 19 August 2026, including direct and indirect participation arrangements.Evidence A · P1
  138. MWITA-SI-2026-026SWIFT ended the MT and ISO 20022 coexistence period for cross-border payment instructions on 22 November 2025 and made ISO 20022 the required instruction language.Evidence A · P1
  139. MWITA-SI-2026-027SWIFT reported that more than 97% of messages on its network were sent in ISO 20022 format after the November 2025 cutover.Evidence B · P1
  140. MWITA-AC-2026-014Brazil's care law requires improved statistical and administrative data to recognize and measure the economic and social value of unpaid care work.Evidence A · P1
  141. MWITA-AC-2026-034WHO's GAPS framework includes households, individuals, caregivers and older people in long-term-care facilities, addressing a common institutional-coverage gap.Evidence A · P1
  142. MWITA-CRM-2026-001USGS Mineral Commodity Summaries 2026 presents reserves and resources under separate classifications and provides dedicated definitions and data-source notes for each.Evidence A · P1
  143. MWITA-CRM-2026-002USGS estimates Indonesia mined 2.6 million tonnes of nickel in 2025, about two-thirds of the 3.9 million-tonne world total.Evidence A · P1
  144. MWITA-CRM-2026-003USGS lists Indonesia's nickel reserves at 62 million tonnes versus a world total above 140 million tonnes.Evidence A · P1
  145. MWITA-CRM-2026-004USGS estimates Australian nickel mine production fell 54% to 45,000 tonnes in 2025 after mines entered care and maintenance because of low prices, while listed reserves remained 25 million tonnes.Evidence A · P1
  146. MWITA-CRM-2026-005USGS estimates the Democratic Republic of the Congo mined 230,000 tonnes of cobalt in 2025, 73% of world mine output.Evidence A · P1
  147. MWITA-CRM-2026-006USGS lists DRC cobalt reserves at 6 million tonnes, world reserves at 12 million tonnes and identified terrestrial world resources at about 25 million tonnes.Evidence A · P1
  148. MWITA-CRM-2026-007USGS estimates Indonesia mined 44,000 tonnes of cobalt in 2025, 14% of world output and second to the DRC.Evidence A · P1
  149. MWITA-CRM-2026-008USGS reports that DRC replaced a temporary cobalt export ban with an 18,125-tonne contained-cobalt quota for the remainder of 2025 and up to 96,600 tonnes annually in 2026 and 2027, including 9,600 tonnes for strategic reserves.Evidence A · P1
  150. MWITA-CRM-2026-009A 2026 USGS study estimates China produced 98% of global primary refined gallium in 2023 and held 91% of annual primary gallium capacity.Evidence A · P1
  151. MWITA-CRM-2026-010The USGS study estimates China produced 154 tonnes of refined germanium in 2023, 77% of global production.Evidence A · P1
  152. MWITA-CRM-2026-011USGS estimates China produced 3.63 million tonnes of silicon metal in 2023, 85% of world output.Evidence A · P1
  153. MWITA-CRM-2026-012China's Ministry of Commerce and customs authority announced export controls on specified medium and heavy rare-earth-related items in Announcement No.18 of 2025.Evidence A · P1
  154. MWITA-CRM-2026-013IEA estimates global battery demand grew more than 35% in 2025 and exceeded 1.5 TWh.Evidence B · P1
  155. MWITA-CRM-2026-014IEA estimates the average share of the leading refiner across key energy minerals reached 70% in 2025, up from 68% in 2020.Evidence B · P1
  156. MWITA-CRM-2026-015IEA estimates Indonesia for nickel and China for most other key energy minerals together accounted for over three-quarters of total refined supply growth from 2023 to 2025.Evidence B · P1
  157. MWITA-CRM-2026-016IEA estimates critical-mineral investment fell 9% in 2025; battery-material company spending fell more than 20% and lithium specialists cut investment around 40%.Evidence B · P1
  158. MWITA-CRM-2026-017IEA's project-pipeline analysis projects a copper supply deficit of about 25% in 2035 under its stated-policy outlook.Evidence C · P1
  159. MWITA-CRM-2026-018IEA reports the leading supplier's rare-earth refining share fell from above 90% in 2023 to 85% in 2025 as US projects and Malaysian output increased.Evidence B · P1
  160. MWITA-CRM-2026-019IEA reports lithium prices more than doubled from early 2025 to early 2026 and cobalt prices rose about 130%, largely after DRC export restrictions.Evidence B · P1
  161. MWITA-CRM-2026-020IEA reports European gallium prices around five times Chinese domestic prices and germanium prices almost three times higher in 2026 amid export controls.Evidence B · P1
  162. MWITA-CRM-2026-021IEA estimates global lithium-ion battery nameplate manufacturing capacity exceeded 4 TWh at end-2025, about 30% above 2024.Evidence B · P1
  163. MWITA-CRM-2026-022IEA estimates China accounted for more than 80% of global lithium-ion nameplate capacity and over 80% of battery output in 2025.Evidence B · P1
  164. MWITA-CRM-2026-023IEA estimates electric-vehicle battery deployment reached 1.2 TWh in 2025, almost 30% above 2024 and more than seven times 2020.Evidence B · P1
  165. MWITA-CRM-2026-024IEA estimates average battery prices declined 8% in 2025, while LFP packs were more than 40% cheaper per kWh than NMC alternatives.Evidence B · P1
  166. MWITA-CRM-2026-025IEA estimates China produced about 85% of global cathode active material and more than 90% of anode active material used in electric-car batteries in 2025.Evidence B · P1
  167. MWITA-CRM-2026-026IEA reports LFP batteries supplied almost half the global electric-car market in 2024-25, up from less than 10% in 2020.Evidence B · P1
  168. MWITA-CRM-2026-027IEA estimates China produces 75% of battery-grade purified phosphoric acid and 95% of high-purity manganese sulphate.Evidence B · P1
  169. MWITA-CRM-2026-028IEA's stated-policy scenario finds announced battery-grade manganese sulphate projects would meet only 55% of expected demand in 2035.Evidence C · P1
  170. MWITA-CRM-2026-029IEA estimates critical minerals account for around one-quarter of battery-cell costs but only about 3% of the price of an average electric vehicle.Evidence B · P1
  171. MWITA-CRM-2026-030South Africa's 2025 Critical Minerals and Metals Strategy calls for a battery value-chain plan and processing hubs while identifying the country as the world's largest manganese producer.Evidence A · P1
  172. MWITA-WH-2026-011WHO states that genotyping choices should account for health-system capacity, ability to follow up women and access to treatment.Evidence A · P1
  173. MWITA-CH-2026-029Participants in the Hadzabe project identified domestic hosting by Tanzanian institutions as important for long-term data management and ownership.Evidence B · P1
  174. MWITA-MM-2026-007World Bank Remittance Prices Worldwide reported a global average quoted cost of 6.36% for the monitored transfer scenario in Q3 2025.Evidence A · P1
  175. MWITA-MM-2026-008For a $200 transfer in Q1 2025, World Bank RPW reported average receiving-market costs of 6.95% for South Africa and 4.98% for Mexico.Evidence A · P1
  176. MWITA-MM-2026-009RPW measures quoted prices in selected corridors; it does not establish which services migrants used or the household outcomes of those transfers.Evidence A · P1
  177. MWITA-MM-2026-010The World Bank estimated officially recorded remittances to low- and middle-income countries at $685 billion for 2024, rather than reporting a final administrative total.Evidence B · P1
  178. MWITA-MM-2026-011Bangko Sentral ng Pilipinas reported $38.341 billion in personal remittances for 2024, 3.0% above 2023.Evidence A · P1
  179. MWITA-MM-2026-012BSP reported $34.493 billion of 2024 cash remittances channeled through banks, below the broader $38.341 billion personal-remittance measure.Evidence A · P1
  180. MWITA-MM-2026-013Bangladesh Bank documented policy measures including a 2.5% cash incentive and simpler transfers alongside stronger formal remittance inflows in FY2024–25.Evidence B · P1
  181. MWITA-MM-2026-014RBI reported India's inward remittances at $118.7 billion in 2023–24, more than double the 2010–11 level.Evidence A · P1
  182. MWITA-MM-2026-015For the common source-country set in RBI's 2023–24 survey, advanced economies accounted for 51.0% of remittances and GCC countries 33.7%.Evidence A · P1
  183. MWITA-AE-2026-001The final IAB/MRC Attention Measurement Guidelines version 1.0 describes four primary approach families: data signals, visual/audio tracking, physiological/neurological observation, and panel- or survey-based methods.Evidence A · P1
  184. MWITA-AE-2026-002The IAB/MRC framework distinguishes viewability, an opportunity for an ad to be seen, from attention, the likelihood that it was noticed or engaged with.Evidence A · P1
  185. MWITA-AE-2026-003The 2025 IAB/MRC framework states that attention is complementary and is not a standalone measure of ad effectiveness or business results.Evidence A · P1
  186. MWITA-AE-2026-004The IAB/CIMM marketer playbook describes attention metrics as probabilistic and advises against treating attention as a binary outcome or universal currency.Evidence A · P1
  187. MWITA-AE-2026-005IAB and IAB Europe’s 2025 commerce-media incrementality guidelines distinguish experiments, model-based counterfactuals, econometric models and hybrid proxies.Evidence A · P1
  188. MWITA-AE-2026-006The 2025 IAB commerce-media guidance centers credible counterfactuals, bias control and separation of signal from noise as three principles for incrementality measurement.Evidence A · P1
  189. MWITA-AE-2026-007The commerce-media guidance places randomized experiments at the strongest direct level for estimating whether advertising caused additional outcomes under the tested conditions.Evidence A · P1
  190. MWITA-AE-2026-008The IAB guidance treats assigning credit to observed sales and estimating sales that would not have occurred without media as different measurement questions.Evidence A · P1
  191. MWITA-AE-2026-009Kuaishou reported RMB81.5 billion in online marketing-services revenue for 2025, up 12.5% from 2024.Evidence A · P1
  192. MWITA-AE-2026-010Kuaishou said online marketing services contributed 57.1% of its RMB142.8 billion total revenue in 2025.Evidence A · P1
  193. MWITA-AE-2026-011Kuaishou management attributed roughly 5% growth in domestic online marketing-services revenue in Q4 2025 to its generative recommendation and intelligent-bidding models.Evidence C · P1
  194. MWITA-AE-2026-012Kuaishou reported RMB4.0 billion of Q4 2025 online-marketing spend associated with AIGC marketing materials.Evidence B · P1
  195. MWITA-AE-2026-013NAVER reported that total platform advertising grew 6.7% year over year in Q4 2025.Evidence A · P1
  196. MWITA-AE-2026-014NAVER reported Q4 2025 Search Platform revenue of KRW1.0596 trillion, down 0.5% year over year, while stating it was up 1.8% excluding LY settlement effects.Evidence A · P1
  197. MWITA-AE-2026-015NAVER reported Q4 2025 Commerce revenue of KRW1.054 trillion, up 36.0% year over year.Evidence A · P1
  198. MWITA-AE-2026-016Sea reported that content-driven orders accounted for more than 20% of Shopee physical-goods order volume in Q2 2025.Evidence B · P1
  199. MWITA-AE-2026-017Sea reported that Shopee’s ad take rate increased by almost 70 basis points year over year in Q2 2025.Evidence B · P1
  200. MWITA-AE-2026-018Sea reported about 20% year-over-year growth in sellers using Shopee ads and more than 40% growth in average quarterly ad spend among ad-paying sellers in Q2 2025.Evidence B · P1
  201. MWITA-AE-2026-019Sea said technology enhancements drove an 8% year-over-year uplift in Shopee purchase-conversion rates in Q2 2025.Evidence C · P1
  202. MWITA-AE-2026-020Sea reported that Shopee ad revenue grew more than 70% year over year in Q4 2025.Evidence A · P1
  203. MWITA-AE-2026-021Sea reported that Shopee’s ad take rate rose by more than 80 basis points year over year in Q4 2025.Evidence B · P1
  204. MWITA-AE-2026-022Sea reported that Shopee ad-paying sellers increased by more than 20% and their average ad spend increased by more than 45% year over year in Q4 2025.Evidence B · P1
  205. MWITA-AE-2026-025A 2025 Taiwan eye-tracking study enrolled 71 university adults who each viewed 12 product-placement video stimuli mixed with 24 filler videos.Evidence A · P1
  206. MWITA-AE-2026-026The Taiwan study found that visual attention to embedded product placements varied with the placement, surrounding story context and viewers’ prior fixation dynamics.Evidence A · P1
  207. MWITA-AE-2026-027In the Taiwan study, verbal references to a product category increased fixation on product placements, with the pattern moderated by story relevance and visual salience.Evidence A · P1
  208. MWITA-AE-2026-028The Taiwan product-placement study’s primary outcome was fixation proportion inside a dynamic area of interest, not clicks, conversions, sales or brand lift.Evidence A · P1
  209. MWITA-AE-2026-029A field experiment with an Indian direct-to-consumer retailer randomized WhatsApp recipients among GenAI-personalized video, personalized image and generic non-personalized video advertising conditions.Evidence A · P1
  210. MWITA-AE-2026-030The India field experiment reported that GenAI-personalized video ads increased engagement by 6–9 percentage points relative to the personalized-image and generic-video baselines.Evidence A · P1
  211. MWITA-AE-2026-031The India experiment’s reported 6–9 percentage-point effect concerns message engagement, so it cannot by itself establish advertiser return or customer welfare.Evidence A · P1
  212. MWITA-AE-2026-032Google Ads reported in November 2025 that its Conversion Lift product had reduced spend and conversion-volume thresholds and could test a manager account or selected campaigns.Evidence C · P1
  213. MWITA-AE-2026-033Google’s 2025 guidance distinguishes incrementality experiments, marketing-mix models and attribution as answering causal campaign, aggregate portfolio and touchpoint-credit questions respectively.Evidence C · P1
  214. MWITA-AE-2026-034Google’s statement that improved lift tools measure true causal impact is a product-method claim, not evidence that any particular advertiser achieved positive lift.Evidence C · P1
  215. MWITA-AE-2026-035Singapore’s PDPC and IMDA launched a PET Adoption Guide in July 2025 based on use cases from their PET Sandbox.Evidence A · P1
  216. MWITA-AE-2026-036Singapore’s 2025 PET guidance announcement says PETs can support cross-organization insight sharing and machine-learning access without directly exposing underlying data.Evidence B · P1
  217. MWITA-AE-2026-037India notified the Digital Personal Data Protection Rules, 2025 on 14 November 2025, with an official enforcement timeline rather than simultaneous operation of every provision.Evidence A · P1
  218. MWITA-AE-2026-038India’s 2025 rules require data-fiduciary notices to be clear and standalone and to itemize personal data and processing purposes relevant to informed consent.Evidence A · P1
  219. MWITA-CSE-2026-001India's HCES estimated 2023-24 average monthly per-capita consumption expenditure excluding imputed free welfare items at ₹4,122 in rural areas and ₹6,996 in urban areas.Evidence A · P1
  220. MWITA-CSE-2026-002Food represented 47.04% of rural MPCE in India's 2023-24 HCES.Evidence A · P1
  221. MWITA-CSE-2026-003Food represented 39.68% of urban MPCE in India's 2023-24 HCES.Evidence A · P1
  222. MWITA-CSE-2026-004Cereals accounted for 4.99% of rural MPCE and 3.76% of urban MPCE in India's 2023-24 HCES.Evidence A · P1
  223. MWITA-CSE-2026-005Non-food expenditure represented approximately 53% of rural and 60% of urban MPCE in India's 2023-24 HCES.Evidence A · P1
  224. MWITA-CSE-2026-006India's official HCES summary reported rural consumption Gini falling from 0.266 to 0.237 and urban Gini from 0.314 to 0.284 between 2022-23 and 2023-24.Evidence A · P1
  225. MWITA-CSE-2026-007The official HCES release reported nominal MPCE growth for the bottom 5% of about 22% in rural India and 19% in urban India between 2022-23 and 2023-24.Evidence A · P1
  226. MWITA-CSE-2026-008India's HCES measures consumption through the value of goods and services acquired for consumption, whether or not payment occurred in the same reference period.Evidence A · P1
  227. MWITA-CSE-2026-009China's 2025 nationwide per-capita disposable income was ¥43,377, up 5.0% nominally and 5.0% in real terms from 2024.Evidence A · P1
  228. MWITA-CSE-2026-010China's 2025 median per-capita disposable income was ¥36,231, 83.5% of the average, and grew 4.4% nominally.Evidence A · P1
  229. MWITA-CSE-2026-011China's 2025 per-capita consumption expenditure was ¥29,476, up 4.4% nominally and 4.4% in real terms.Evidence A · P1
  230. MWITA-CSE-2026-012In 2025 China's per-capita consumption expenditure was ¥35,869 in urban areas and ¥20,259 in rural areas; nominal growth was 3.8% and 5.1%, respectively.Evidence A · P1
  231. MWITA-CSE-2026-013Food, tobacco and alcohol represented 29.3% of China's per-capita consumption expenditure in 2025, down 0.5 percentage point; the shares were 28.3% urban and 31.8% rural.Evidence A · P1
  232. MWITA-CSE-2026-014China's 2025 per-capita service consumption expenditure was ¥13,602, up 4.5%, and accounted for 46.1% of total expenditure, unchanged from 2024.Evidence A · P1
  233. MWITA-CSE-2026-015China's per-capita expenditure on household facilities, articles and services rose 7.7% nominally in 2025.Evidence A · P1
  234. MWITA-CSE-2026-016China's per-capita transport and communications expenditure rose 8.3% nominally in 2025.Evidence A · P1
  235. MWITA-CSE-2026-017China's per-capita education, culture and recreation expenditure rose 9.4% nominally in 2025.Evidence A · P1
  236. MWITA-CSE-2026-018China's per-capita healthcare expenditure rose 1.0% nominally in 2025.Evidence A · P1
  237. MWITA-CSE-2026-019China's 2025 per-capita disposable income ranged from ¥10,150 in the low-income group to ¥103,778 in the high-income group in NBS's five-group distribution.Evidence A · P1
  238. MWITA-CSE-2026-020Mexico's ENIGH 2024 reported average monthly current household income of MXN25,955, 10.6% above 2022.Evidence A · P1
  239. MWITA-CSE-2026-021Mexico's ENIGH 2024 reported average monthly current monetary household spending of MXN15,891.Evidence A · P1
  240. MWITA-CSE-2026-022In Mexico's ENIGH 2024, food, beverages and tobacco represented 37.7% of current monetary household spending and transport and communications 19.5%.Evidence A · P1
  241. MWITA-CSE-2026-023Mexico's ENIGH 2024 reported average quarterly household health spending of MXN2,424, equal to 3.7% of current monetary spending.Evidence A · P1
  242. MWITA-CSE-2026-024ENIGH 2024 reported men's average monthly monetary income was MXN4,111 higher than women's in Mexico.Evidence A · P1
  243. MWITA-CSE-2026-025ENIGH 2024 reported average income among people self-identifying as Indigenous or speaking an Indigenous language was 26.0% below Mexico's national average.Evidence A · P1
  244. MWITA-CSE-2026-026Mexico's ENIGH 2024 used a probabilistic multistage stratified cluster sample of 105,718 dwellings and reported 1.60% relative error for national quarterly household current income.Evidence A · P1
  245. MWITA-CSE-2026-027Brazil's 2025 national accounts recorded household consumption expenditure of R$8.1 trillion.Evidence A · P1
  246. MWITA-CSE-2026-028Brazil's household consumption expenditure was flat quarter-on-quarter in Q4 2025 after seasonal adjustment and 1.0% above Q4 2024.Evidence A · P1
  247. MWITA-CSE-2026-029Brazil's nominal monthly per-capita household income was R$2,316 in 2025.Evidence A · P1
  248. MWITA-CSE-2026-030Brazil's 2025 nominal monthly per-capita household income ranged from R$1,219 in Maranhão to R$4,538 in the Federal District.Evidence A · P1
  249. MWITA-CSE-2026-031Brazil's official IPCA consumer-price index rose 4.26% over 2025.Evidence A · P1
  250. MWITA-CSE-2026-032Japan's two-or-more-person households spent an average ¥314,001 per month in 2025, up 4.6% nominally and 0.9% in real terms.Evidence A · P1
  251. MWITA-CSE-2026-033Japan's worker households with two or more persons had average monthly income of ¥653,901 in 2025, up 2.8% nominally but down 0.9% in real terms.Evidence A · P1
  252. MWITA-CSE-2026-034Japan's worker households with two or more persons averaged ¥346,297 monthly consumption expenditure in 2025, up 6.5% nominally and 2.7% in real terms.Evidence A · P1
  253. MWITA-CSE-2026-035Japan's Statistics Bureau reported a 0.4% standard-error rate for the 2025 yearly average consumption expenditure of two-or-more-person households, with about 7,261 tabulated households.Evidence A · P1
  254. MWITA-CSE-2026-036South Korean households averaged KRW2.950 million monthly consumption expenditure in Q1 2025, up 1.4% nominally but down 0.7% at 2020 prices year-on-year.Evidence A · P1
  255. MWITA-CSE-2026-037South Korean average monthly household income was KRW5.351 million in Q1 2025, up 4.5% nominally and 2.3% at 2020 prices year-on-year.Evidence A · P1
  256. MWITA-CSE-2026-038South Korean household expenditure on housing, water, electricity and fuels rose 5.8% nominally and 3.8% in real terms year-on-year in Q1 2025.Evidence A · P1
  257. MWITA-CSE-2026-039In South Korea Q1 2025 household transportation spending fell 3.7% nominally and 5.8% in real terms, while clothing and footwear fell 4.7% nominally and 6.4% in real terms year-on-year.Evidence A · P1
  258. MWITA-CSE-2026-040South Korea's average propensity to consume was 69.8% in Q1 2025, 2.1 percentage points below Q1 2024.Evidence A · P1
  259. MWITA-CSE-2026-041In South Korea Q1 2025 the lowest income quintile's average propensity to consume was 147.6%, versus 56.7% for the highest quintile.Evidence A · P1
  260. MWITA-CSE-2026-042Indonesia's economy grew 5.11% in 2025, with GDP per capita at current prices of IDR83.7 million.Evidence A · P1
  261. MWITA-WS-2026-006The World Bank estimates that cities and industries generate about one billion cubic metres of used water each day.Evidence A · P1
  262. MWITA-WS-2026-007The World Bank estimates current water reuse at about 12 percent of municipal freshwater withdrawals, while potable and industrial reuse accounts for about 3 percent.Evidence A · P1
  263. MWITA-WS-2026-009The World Bank models potential reuse of up to 430 million cubic metres per day by 2040 and an associated investment need of up to US$340 billion.Evidence A · P1
  264. MWITA-WS-2026-022India states a central Jal Jeevan Mission outlay of ₹2,08,652 crore for the programme period cited in its December 2025 release.Evidence A · P1
  265. MWITA-WS-2026-026Australia's Murray–Darling Basin framework permits trading of water rights and government purchases of entitlements for environmental purposes.Evidence A · P1
  266. MWITA-WS-2026-027Australian government water-market price reports describe observed transactions but explicitly are not financial advice.Evidence A · P1
  267. MWITA-WS-2026-029Brazil's water-resources law includes permits and charging for water use among its policy instruments.Evidence A · P1
  268. MWITA-WS-2026-032The Mohamed bin Zayed Water Initiative and World Bank signed a 2025 memorandum to explore water innovation and investment collaboration.Evidence A · P1
  269. MWITA-WS-2026-048Saudi Water Partnership Company projected its desalination portfolio rising from 4.16 million cubic metres per day in 2024 to about 7.37 million in 2030.Evidence A · P1
  270. MWITA-WS-2026-049SWPC's seven-year statement listed seven planned desalination plants totaling 2.75 million cubic metres per day.Evidence A · P1
  271. MWITA-WS-2026-052India's Jal Hi Amrit initiative aims to incentivize sustained operation of treatment plants producing recyclable water that meets environmental standards.Evidence A · P1
  272. MWITA-II-2026-001IAIS reports differing 2024 cyber-insurance market estimates: USD 15.3 billion from Munich Re and USD 14.7 billion from Swiss Re.Evidence A · P1
  273. MWITA-II-2026-002More than 90 percent of affirmative cyber premiums in IAIS's monitored sample were written or assumed by insurer groups headquartered in Europe or North America.Evidence A · P1
  274. MWITA-II-2026-003For 23 insurers consistently reporting to IAIS, affirmative cyber gross written premiums rose only modestly between year-end 2021 and year-end 2024 while net incurred cyber claims declined.Evidence A · P1
  275. MWITA-II-2026-005IAIS concludes that natural-catastrophe protection gaps remain significant globally and are especially consequential in emerging and developing economies.Evidence A · P1
  276. MWITA-II-2026-006IAIS's historical case studies did not find financial instability from the examined disasters, while its analysis identifies widening gaps as a possible channel shifting risk toward banks and public balance sheets.Evidence A · P1
  277. MWITA-II-2026-011OECD's 2025 insurance report covers 67 jurisdictions, including 14 non-OECD Latin American and 15 other Asian or European jurisdictions.Evidence A · P1
  278. MWITA-II-2026-012OECD reports that insurance penetration rose in 2024 but remained below its level ten years earlier.Evidence A · P1
  279. MWITA-II-2026-013Across OECD reporting jurisdictions, non-life gross premiums grew on average by 8.2 percent nominally and 4.2 percent in real terms in 2024, while claims payments grew 7.5 percent nominally and 3.5 percent in real terms.Evidence A · P1
  280. MWITA-II-2026-014OECD reports a combined ratio below 100 percent in almost all reporting jurisdictions and improvement from 2023 in 72 percent of reporting jurisdictions.Evidence A · P1
  281. MWITA-II-2026-015Swiss Re estimated global insured natural-catastrophe losses at USD 137 billion in 2024.Evidence A · P1
  282. MWITA-II-2026-016Swiss Re estimated that 57 percent of 2024 disaster economic losses were uninsured, leaving a USD 181 billion global natural-catastrophe protection gap.Evidence A · P1
  283. MWITA-II-2026-017Swiss Re estimated 2025 natural-catastrophe economic losses at USD 220 billion, with about 49 percent insured.Evidence A · P1
  284. MWITA-II-2026-018Although the insured share of 2025 natural-catastrophe losses was the highest in Swiss Re's series, more than half of estimated losses remained uninsured.Evidence A · P1
  285. MWITA-II-2026-019Munich Re estimated global cyber-insurance premium volume at USD 15.3 billion in 2024.Evidence A · P1
  286. MWITA-II-2026-020Munich Re estimated North America accounted for USD 10.6 billion, or 69 percent, of global cyber-insurance premiums in 2024.Evidence A · P1
  287. MWITA-II-2026-021Munich Re projected global cyber-insurance premiums of USD 16.3 billion for 2025.Evidence A · P1
  288. MWITA-II-2026-027ARC presented just over USD 5.4 million in parametric payouts to Mozambique and the World Food Programme in October 2025 for drought and Cyclone Chido response.Evidence A · P1
  289. MWITA-II-2026-028The Mozambique payout comprised just over USD 1.8 million to government for drought response and USD 3.5 million to WFP across drought and cyclone policies.Evidence A · P1
  290. MWITA-II-2026-029WFP's Cyclone Chido policy component released USD 3.1 million for response in Nampula and Cabo Delgado.Evidence A · P1
  291. MWITA-II-2026-030Malawi received USD 3,376,783 from ARC in July 2025 after drought in the 2024/25 agricultural season.Evidence A · P1
  292. MWITA-II-2026-031Malawi's ARC total combined a USD 3,065,746 traditional sovereign-policy payout and a USD 311,037 anticipatory-policy payout.Evidence A · P1
  293. MWITA-II-2026-033Madagascar became the first subscriber to ARC's sovereign parametric flood policy when the contract was signed in January 2026.Evidence A · P1
  294. MWITA-II-2026-034ARC designed the Madagascar flood mechanism to disburse within up to ten days after a confirmed event crosses the predefined trigger.Evidence A · P1
  295. MWITA-II-2026-035ARC reports issuing more than 300 sovereign policies and paying over USD 250 million in claims since 2014.Evidence A · P1
  296. MWITA-II-2026-036South Africa's National Treasury reported in August 2025 that most public infrastructure was uninsured, creating a government contingent liability.Evidence A · P1
  297. MWITA-II-2026-037South Africa reported that municipal insurance-pool coverage was limited partly by poor asset data and maintenance records.Evidence A · P1
  298. MWITA-II-2026-038South Africa's Treasury and municipalities planned a parametric-insurance pricing and structure pilot by the end of the third quarter of 2025.Evidence A · P1
  299. MWITA-II-2026-039IRDAI published its Handbook on Indian Insurance Statistics 2024-25 catalogue entry on 3 February 2026.Evidence A · P1
  300. MWITA-II-2026-040IRDAI introduced Bima-ASBA in February 2025 as a one-time UPI mandate for blocking premium funds during life and health policy issuance.Evidence A · P1
  301. MWITA-FF-2026-018Shanghai's 2026–2030 food-technology plan targets an industrial output value above RMB 150 billion by 2030.Evidence A · P1
  302. MWITA-FF-2026-019Shanghai targets 8–10 leading niche food-tech firms and one or two synthetic-biology food-company listings by 2030.Evidence A · P1
  303. MWITA-FF-2026-033In three-dimensional microcarrier culture, the same serum-free formulations cost 4.5 to 5.6 times more per million cells than the fetal-bovine-serum comparator.Evidence A · P1
  304. MWITA-FF-2026-035GFI's 2025 review found only four published techno-economic models focused on precision-fermented food proteins.Evidence A · P1
  305. MWITA-FF-2026-036Published precision-fermentation protein models reviewed by GFI spanned roughly four orders of magnitude in modeled cost of production.Evidence A · P1
  306. MWITA-FF-2026-037The GFI review warns that enzyme models omitting downstream concentration and purification do not represent complete precision-fermented food-protein costs.Evidence A · P1
  307. MWITA-FF-2026-038GFI's cultivated-meat review states that published techno-economic models largely represent hypothetical scaled processes rather than observed commercial plants.Evidence A · P1
  308. MWITA-FF-2026-039The review concludes that the most economically viable cultivated-meat manufacturing method for a given product is not yet known.Evidence A · P1
  309. MWITA-FF-2026-040The review identifies static stoichiometry, homogeneous mixing and fixed feeding or growth assumptions as limits on existing cultivated-meat models.Evidence A · P1
  310. MWITA-FF-2026-041Solar Foods reported that Factory 01 reached its design capacity of 160 tonnes of Solein per year during 2025.Evidence A · P1
  311. MWITA-FF-2026-043Solar Foods reported four LOIs or MOUs covering 6.5–7.65 kilotonnes of prospective annual volume, alongside a planned 6.4-kilotonne Factory 02.Evidence A · P1
  312. MWITA-AD-2026-001Baidu reported that more than half of its mobile search result pages contained AI-generated content at the end of June 2025.Evidence A · P1
  313. MWITA-AD-2026-002Baidu reported that the share of mobile search result pages containing AI-generated content rose from 35% in April 2025 to 64% in July 2025.Evidence A · P1
  314. MWITA-AD-2026-003Baidu reported 735 million monthly active users for the Baidu App in June 2025, 5% more than a year earlier.Evidence A · P1
  315. MWITA-AD-2026-004Baidu's online marketing revenue was RMB16.2 billion in Q2 2025, down 15% year over year, while non-online-marketing revenue rose 34% to RMB10.0 billion.Evidence A · P1
  316. MWITA-AD-2026-005Alibaba launched a Qwen-powered version of Quark on 13 March 2025 as an all-in-one AI assistant with chat, reasoning, research and task-execution functions.Evidence B · P1
  317. MWITA-AD-2026-006Alibaba described the pre-existing Quark information-services platform as serving more than 200 million users in China when it announced the upgraded AI assistant.Evidence B · P1
  318. MWITA-AD-2026-007Alibaba stated that Quark's multi-source answers include links to references.Evidence B · P1
  319. MWITA-AD-2026-008NAVER launched AI Briefing on 27 March 2025 in desktop and mobile integrated search without a separate subscription signup.Evidence B · P1
  320. MWITA-AD-2026-009At launch, NAVER AI Briefing supported official or multi-source summaries, short-form content, places and shopping formats.Evidence B · P1
  321. MWITA-AD-2026-010NAVER framed connecting 40 million users with content and increasing exposure opportunities for 3.3 million SMEs as aims of AI Briefing, not measured outcomes.Evidence B · P1
  322. MWITA-AD-2026-011NAVER reported in June 2025 that AI Briefing exposure was about three times its launch level.Evidence B · P1
  323. MWITA-AD-2026-012NAVER reported that the click-through rate on AI Briefing's 'more' button was 50% higher than at launch.Evidence B · P1
  324. MWITA-AD-2026-013NAVER reported that related-question click-through in AI Briefing was 3.4 times that of its existing search-query recommendation area.Evidence B · P1
  325. MWITA-AD-2026-014NAVER reported AI Briefing area CTR eight percentage points above its existing direct-answer content.Evidence B · P1
  326. MWITA-AD-2026-015NAVER reported a 22% increase in dwell time in the top search-results area after AI Briefing launched.Evidence B · P1
  327. MWITA-AD-2026-016LY Corporation said AI-generated queries represented a mid-single-digit percentage of Yahoo! JAPAN Search queries in its FY2025 Q1 results briefing.Evidence B · P1
  328. MWITA-AD-2026-017LY Corporation said most Yahoo! JAPAN AI-search usage at the FY2025 Q1 briefing involved how-to queries.Evidence B · P1
  329. MWITA-AD-2026-018LY Corporation said monetization methods for Yahoo! JAPAN AI search were undecided at its FY2025 Q1 briefing.Evidence B · P1
  330. MWITA-AD-2026-019Google began rolling AI Mode out to all users in India in English on 8 July 2025 without requiring Labs enrollment.Evidence B · P1
  331. MWITA-AD-2026-020The India rollout of Google AI Mode allowed text, voice and image input and follow-up questions.Evidence B · P1
  332. MWITA-AD-2026-021Google began rolling AI Mode in Hindi out globally on 8 September 2025 after its English launch in India.Evidence B · P1
  333. MWITA-AD-2026-022Google described Hindi AI Mode as using a custom Gemini 2.5 version and supporting text, voice and image queries.Evidence B · P1
  334. MWITA-AD-2026-023Google states that a page must be indexed and eligible for a search-result snippet to be eligible as a supporting link in AI Overviews or AI Mode.Evidence A · P1
  335. MWITA-AD-2026-024Google says there are no additional technical requirements or special markup needed for inclusion in AI Overviews or AI Mode beyond foundational Search requirements.Evidence A · P1
  336. MWITA-AD-2026-025Google states that AI Overviews and AI Mode traffic is included within the Search Console Performance report under the Web search type.Evidence A · P1
  337. MWITA-AD-2026-026Microsoft launched AI Performance in Bing Webmaster Tools as a public preview on 10 February 2026.Evidence B · P1
  338. MWITA-AD-2026-027Bing AI Performance reports total citations, average cited pages, sampled grounding-query phrases and page-level citation activity across supported AI surfaces.Evidence A · P1
  339. MWITA-AD-2026-028Microsoft says Bing AI Performance reflects robots.txt and other supported content-owner preferences.Evidence A · P1
  340. MWITA-AD-2026-029Cloudflare estimated June 2025 crawl-to-referral ratios of about 14:1 for Google, 1,700:1 for OpenAI and 73,000:1 for Anthropic on its network.Evidence A · P1
  341. MWITA-AD-2026-031Cloudflare explicitly cautioned that crawl-to-referral ratios may undercount referrals from native apps when no Referer header is sent.Evidence A · P1
  342. MWITA-AD-2026-032Adobe Analytics observed that traffic from generative-AI sources to U.S. retail sites was 4,700% higher year over year in July 2025.Evidence A · P1
  343. MWITA-AD-2026-033In July 2025 Adobe measured AI-referred U.S. retail visits as 32% longer, with 10% more pages per visit and a 27% lower bounce rate than non-AI traffic sources.Evidence A · P1
  344. MWITA-AD-2026-034Adobe measured AI-referred U.S. retail traffic as 23% less likely to convert than non-AI traffic in July 2025, versus a 49% deficit in January 2025.Evidence A · P1
  345. MWITA-AD-2026-035OpenAI updated ChatGPT search on 5 February 2025 to make it available without signup in regions where ChatGPT is available.Evidence B · P1
  346. MWITA-AD-2026-036ChatGPT search answers can include inline links and a Sources panel linking to web sources.Evidence B · P1
  347. MWITA-AD-2026-037The Tow Center tested 1,600 article-identification queries across eight generative search tools in 2025.Evidence A · P1
  348. MWITA-AD-2026-038The Tow Center audit found fabricated links, citations to syndicated or copied articles, and incorrect or speculative responses when tools could not identify the requested news article.Evidence A · P1
  349. MWITA-AD-2026-039The GEO study introduced a 10,000-query benchmark and reported content-visibility gains of up to 40% from tested optimization methods.Evidence A · P1
  350. MWITA-AD-2026-040The GEO study found optimization effectiveness varied by query domain and that keyword stuffing reduced its Perplexity visibility baseline while citations, statistics or quotations improved selected visibility metrics.Evidence A · P1
  351. MWITA-SC-2026-002India reported that 237 projects had accessed subsidized AI computing by August 2026.Evidence A · P1
  352. MWITA-SC-2026-003India reported cumulative subsidized usage of 9.318 million GPU-hours by August 2026.Evidence A · P1
  353. MWITA-SC-2026-004India reported selecting 20 indigenous foundation-model proposals from 506 applications by August 2026.Evidence A · P1
  354. MWITA-SC-2026-034The Korean centre is intended to support industry, academia and research and promote domestic AI semiconductors.Evidence B · P1
  355. MWITA-ASVC-2026-052stc launched a voice-of-customer transcription co-worker using its speech-recognition and large-language-model capabilities.Evidence B · P1
  356. MWITA-ASVC-2026-053stc says its call-transcription co-worker extracts signals for churn, CSAT and fraud analysis.Evidence B · P1
  357. MWITA-SR-2026-001GPT-3 conditioned on thousands of sociodemographic backstories from real US survey participants produced response patterns the authors described as algorithmic fidelity.Evidence A · P0
  358. MWITA-SR-2026-002The study's evidence is fine-grained and demographically conditioned rather than evidence that one unconditioned model represents everyone.Evidence A · P0
  359. MWITA-SR-2026-003ChatGPT synthetic feeling-thermometer means were close to ANES means while response dispersion was lower than in the human surveys.Evidence A · P0
  360. MWITA-SR-2026-004Regression coefficients from synthetic responses often differed significantly from estimates using the human ANES sample.Evidence A · P0
  361. MWITA-SR-2026-005The benchmark found exaggerated ideological out-group antipathy in some persona-conditioned outputs.Evidence A · P0
  362. MWITA-SR-2026-006Turing Experiments replicated the established direction in the Ultimatum Game, Garden Path and Milgram studies for tested recent models.Evidence A · P0
  363. MWITA-SR-2026-007The Wisdom of Crowds Turing Experiment exposed a hyper-accuracy distortion in some models including GPT-4.Evidence A · P0
  364. MWITA-SR-2026-008Agents built from qualitative interviews with 1,052 individuals were evaluated against those same individuals across surveys, traits and experiments.Evidence B · P0
  365. MWITA-SR-2026-009The agents matched GSS responses 85% as accurately as participants matched their own answers two weeks later.Evidence B · P0
  366. MWITA-SR-2026-010Interview-grounded agents reduced racial and ideological accuracy bias compared with agents given only demographic descriptions.Evidence B · P0
  367. MWITA-SR-2026-011The global-distribution benchmark used the 2017–2022 World Values Survey covering 66 countries and more than 80,000 human respondents.Evidence A · P0
  368. MWITA-SR-2026-012Fine-tuning on response distributions outperformed tested prompting and zero-shot baselines on seen and unseen benchmark splits.Evidence A · P0
  369. MWITA-SR-2026-013Even the best specialized models remained far from perfect, particularly on unseen questions.Evidence A · P0
  370. MWITA-SR-2026-014Tested LLMs produced less cross-country diversity than the human survey data.Evidence A · P0
  371. MWITA-SR-2026-015GlobalOpinionQA was built from cross-national survey questions and answers to compare model outputs with country-conditioned human distributions.Evidence B · P0
  372. MWITA-SR-2026-016Default model responses were closer to opinions in the United States and some European and South American populations than to all populations equally.Evidence B · P0
  373. MWITA-SR-2026-017Country prompting shifted outputs toward the prompted population but could also produce harmful cultural stereotypes.Evidence B · P0
  374. MWITA-SR-2026-018Translating questions into a target language did not necessarily make model responses most similar to speakers of that language.Evidence B · P0
  375. MWITA-SR-2026-019Across the benchmark's four models, no LLM beat the strongest non-LLM baseline for individual-level prediction.Evidence B · P0
  376. MWITA-SR-2026-020The tested models over-determined demographics, making identity far more predictive of attitudes than in real respondents.Evidence B · P0
  377. MWITA-SR-2026-021On the study's targeting task, models inflated between-segment gaps two- to fourfold.Evidence B · P0
  378. MWITA-SR-2026-022The synthetic targeting exercise selected the wrong segment in half of US cases and most cross-cultural cases in that benchmark.Evidence B · P0
  379. MWITA-SR-2026-023The Philippine sentiment-simulation framework instantiated agents from a nationally representative survey of 2,485 respondents.Evidence B · P0
  380. MWITA-SR-2026-024The framework separates agent embodiment, scenario exposure and sentiment generation with rationales.Evidence B · P0
  381. MWITA-SR-2026-025A China–US comparison evaluated DeepSeek, Qwen, GPT and Llama families on ANES and China's Zuobiao public-opinion data.Evidence B · P0
  382. MWITA-SR-2026-026A preliminary experiment tested whether LLMs reproduce personality differences between US and South Korean targets.Evidence B · P0
  383. MWITA-SR-2026-027Google DeepMind's Concordia implements entities, modular components, simulation engines and Game Master-mediated environments for generative social simulation.Evidence B · P0
  384. MWITA-SR-2026-028Concordia includes questionnaire, open-ended interview, marketplace and psychology-experiment components.Evidence B · P0
  385. MWITA-SR-2026-029Synthetic Users documents audience definitions spanning demographics, psychographics, behaviors and context, with AI-generated or extended personas.Evidence C · P0
  386. MWITA-SR-2026-030The platform documents structured or semi-structured synthetic interviews, dynamic follow-ups and full transcripts.Evidence C · P0
  387. MWITA-SR-2026-031The platform's analysis flow converts synthetic transcripts into summaries, knowledge graphs, reports and follow-up analysis.Evidence C · P0
  388. MWITA-SR-2026-032Qualtrics stated in February 2026 that its generally available synthetic panel covered the US general population in English.Evidence C · P0
  389. MWITA-SR-2026-033Qualtrics frames synthetic responses as useful for early-stage exploration and hypothesis testing.Evidence C · P0
  390. MWITA-SR-2026-035Kantar states LINK AI predicts ad-effectiveness scores in as little as 15 minutes without a consumer sample.Evidence C · P0
  391. MWITA-SR-2026-036LINK AI documentation says its training base includes more than 230,000 tests, over ten years of data and more than 30 million human interactions.Evidence C · P0
  392. MWITA-SR-2026-038Ipsos described a Ground-Truth Database initiative with Stanford researchers to develop and validate digital twins.Evidence C · P0
  393. MWITA-SR-2026-039Ipsos PersonaBots are presented as a way to converse with existing research segments about concepts, messages and competitive scenarios.Evidence C · P0
  394. MWITA-SR-2026-042ESOMAR asks buyers to verify how training datasets were selected, their coverage, recency and representativeness for the deployment environment.Evidence B · P0
  395. MWITA-SR-2026-043ESOMAR asks what validation compared synthetic outputs with primary research or real-world results.Evidence B · P0
  396. MWITA-SR-2026-044ESOMAR asks suppliers to flag and distinguish data derived from natural persons from data derived synthetically.Evidence B · P0
  397. MWITA-SR-2026-047ICO guidance warns that bias in the original data can be carried into synthetic data.Evidence A · P0
  398. MWITA-SR-2026-055The Chinese measures require steps to prevent discrimination across ethnicity, belief, nationality, region, sex, age, occupation and health.Evidence A · P0
  399. MWITA-SR-2026-056The Chinese measures call for transparency and improved accuracy and reliability of generated content.Evidence A · P0
  400. MWITA-SR-2026-062ANPD said the first sandbox cycle structured test environments and did not yet fully test the solutions.Evidence A · P0
  401. MWITA-SR-2026-063The first ANPD report found that synthetic-data protocols needed improvement.Evidence A · P0
  402. MWITA-SR-2026-066A defensible procurement benchmark must disclose the human comparator, sampling frame, split logic, model version, prompt, geography, language and error distribution.Evidence C · P0
  403. MWITA-SR-2026-067For Metatron, transcript realism, thematic overlap and persona consistency must remain secondary to held-out decision error and subgroup calibration.Evidence C · P0
  404. MWITA-SR-2026-068Temporal validation is required because a model calibrated to a past survey wave can drift from current preferences even if its original benchmark passed.Evidence C · P0
  405. MWITA-SR-2026-070Automated creative scores should triage what receives scarce human testing, not silently replace market-specific validation for high-spend or high-risk campaigns.Evidence C · P0
  406. MWITA-SR-2026-071Synthetic interview transcripts and quotes must be visibly labeled because anthropomorphic language can make generated testimony look like consented lived experience.Evidence C · P0
  407. MWITA-SR-2026-072Non-Western validity requires local human anchors, languages and research partners; geographic persona labels alone are insufficient.Evidence C · P0
  408. MWITA-SR-2026-073A release gate for synthetic respondents should independently score privacy leakage, aggregate distribution fit, individual prediction, subgroup calibration, temporal stability and decision regret.Evidence C · P0
  409. MWITA-SR-2026-074Synthetic participant outputs can be useful as falsifiable hypotheses, adversarial edge cases and interview-script tests even when they are invalid as population estimates.Evidence C · P0
  410. MWITA-SR-2026-075No source in this bundle demonstrates an independently audited commercial decision outcome caused by synthetic-consumer evidence.Evidence D · P0
  411. MWITA-NA-2026-003A laboratory study collected EEG and eye movements from Chinese, German and French participants and found culture-related differences in model performance and neural patterns.Evidence B · P1
  412. MWITA-NA-2026-019India's Department of Biotechnology issued a call seeking research on neural mechanisms, neurotool interfaces and resulting proofs of concept.Evidence A · P0
  413. MWITA-NA-2026-021An Indian-authored study compared machine-learning classification of affective levels and basic emotions using EEG, ECG and GSR signals from the AMIGOS and ASCERTAIN datasets.Evidence B · P1
  414. MWITA-NA-2026-029Researchers at Lagos State University evaluated an ensemble bagged-tree classifier for EEG emotion recognition using the public DEAP dataset.Evidence B · P1
  415. MWITA-MO-2026-001In a randomized trial with 640 Kenyan entrepreneurs, access to a GPT-4 WhatsApp mentor produced no positive average revenue or profit effect; estimated effects were about +15% for higher-baseline performers and about -8% to -10% for lower-baseline performers.Evidence A · P1
  416. MWITA-MO-2026-002A randomized field experiment in meal-delivery support found that generative-AI assistance shortened response time and improved customer sentiment, especially for less-experienced agents, while rapid AI-assisted replies after a failed bot handoff could reduce sentiment.Evidence A · P1
  417. MWITA-MO-2026-003In a two-week cluster-randomized experiment across 16 Sam's Club stores, a generative-AI ideation tool increased promoted-product revenue by 9.76% on average, with a 31.35% estimate for frontline managers and 2.46% for top managers.Evidence B · P1
  418. MWITA-MO-2026-004In a preregistered financial-regulator experiment, generative AI improved document-task quality by 17% and completion time by 34%, but reduced data-task quality by 12% and did not significantly change mean data-task completion time.Evidence A · P1
  419. MWITA-MO-2026-005In a 10-week Facebook A/B test spanning nearly 35,000 advertisers and 640,000 ad variations, AdLlama increased click-through rate by 6.7% versus a supervised imitation model.Evidence A · P1
  420. MWITA-PE-2026-001Adding GPT-extracted listing-text features improved an Oslo XGBoost AVM from 10.53% to 7.97% RMSE and from 6.36% to 5.39% MAPE.Evidence A · P1
  421. MWITA-SEC-2026-003PoisonedRAG reached up to 90% attack success by inserting five malicious texts per target query into knowledge bases containing millions of texts.Evidence A · P1
  422. MWITA-AGC-2026-008On complex live-web shopping queries, evaluated web and deep-research agents achieved low holistic success, especially when filters and sorting had to be satisfied together.Evidence B · P1
  423. MWITA-HCO-2026-009A minimal prompting boost improved the evidence-based quality of screening information produced by three LLMs for lay users, but responses remained below required health-communication standards.Evidence A · P1
  424. MWITA-HCO-2026-012In a randomized simulated-consultation crossover study, AMIE exceeded primary-care physicians on diagnostic accuracy and most rated communication dimensions, but the unfamiliar text-only OSCE setting prevents inference to clinical deployment.Evidence A · P1
  425. MWITA-AMPO-2026-001LLM and hybrid email copy produced gross profit statistically indistinguishable from human copy while sharply reducing production time in three retailer field experiments.Evidence A · P1
  426. MWITA-AMPO-2026-002A RAG-LLM customer-service bot increased helpful feedback and modestly reduced human escalation, while a legacy flow bot remained better for simple operations.Evidence B · P1
  427. MWITA-AMPO-2026-003Offering AI-generated video titles increased platform views and watch time at scale, but substituting for a title a creator would have written could reduce reach.Evidence B · P1
  428. MWITA-AMPO-2026-004An AI-first bank sales funnel increased loan applications only when human agents performed follow-up using AI intent scores.Evidence B · P1
  429. MWITA-AMPO-2026-005AI outbound calls had no significant average effect on resuming account opening and reduced intent through lower perceived empathy in follow-up evidence.Evidence A · P1
  430. MWITA-AMPO-2026-006Human–AI collaboration improved creativity in a short ideation task, but did not outperform human–human collaboration and varied nonlinearly by baseline ability.Evidence A · P1
  431. MWITA-AMPO-2026-007An AI digital streamer increased retail sales only when given interactive Q&A or lottery functionality; the basic streamer did not.Evidence A · P1
  432. MWITA-AMPO-2026-008Showing startups how peers reorganized production around AI increased use-case discovery, task completion, paying customers and reported revenue without reducing labor demand.Evidence B · P1
  433. MWITA-AMPO-2026-009The productivity and quality contribution of supplied GPT-4 content differed materially across English, Arabic and Chinese business tasks.Evidence A · P1
  434. MWITA-AMPO-2026-010A Claude-run real office shop failed to make a profit and exhibited pricing, memory, hallucination and control failures.Evidence C · P1
  435. MWITA-AMPO-2026-011A stronger agent produced more and better-priced deals in a real-money employee marketplace, but did not improve satisfaction or perceived fairness.Evidence C · P1
  436. MWITA-AMPO-2026-012Human–AI teams produced substantially more ads and better copy, but worse images, less diversity and no significant live-platform outcome lift.Evidence B · P1
  437. MWITA-LPPO-2026-004Conversational data collection and reciprocal disclosure greatly increased verified personal-information sharing while also raising privacy concern.Evidence A · P1
  438. MWITA-LPPO-2026-012Recommendations tailored to ideal rather than current preferences improved value, trust and reuse intentions despite attracting slightly fewer clicks.Evidence A · P1
  439. MWITA-CCO-2026-003Human sellers attracted more initial interest, whereas automated sellers obtained contact information more readily, showing that the preferred agent type reversed across consecutive funnel stages.Evidence A · P1
  440. MWITA-CCO-2026-004Disclosing possible human involvement behind a hybrid service chatbot made customer messages more human-oriented and increased employee workload.Evidence A · P1
  441. MWITA-CCO-2026-012Automated agents were perceived somewhat less favorably than humans, yet behavioral outcomes such as choosing and buying were statistically equivalent overall.Evidence A · P1
  442. MWITA-PRO-2026-001Price discounts mitigated the post-pandemic food inflation burden, while switching toward cheaper brands partly reversed that relief because cheaper varieties inflated faster.Evidence B · P1
  443. MWITA-PRO-2026-002Experiment-informed competitive price optimization materially increased both revenue and profit at a large US e-commerce retailer.Evidence A · P1
  444. MWITA-PRO-2026-003Providing small firms with readily available competitor prices caused them to change and align their own prices more often.Evidence A · P1
  445. MWITA-PRO-2026-004Larger preregistered replications found no robust purchase-likelihood advantage for multipack prices divisible by unit count.Evidence A · P1
  446. MWITA-PRO-2026-005Spreading the cost of an adaptive innovation across all customers made its pricing more acceptable than charging only the disability-targeted segment a larger premium.Evidence A · P1
  447. MWITA-PRO-2026-006Receiving a low-value outcome from a probabilistic promotion reduced actual repeat purchase relative to receiving the same discount with certainty.Evidence A · P1
  448. MWITA-PRO-2026-007Price-equivalent multi-unit promotions increased units purchased relative to single-unit discounts, but the added items were less likely to be consumed.Evidence A · P1
  449. MWITA-PRO-2026-008Extending a Freemium SaaS trial from three to seven days increased trial adoption and delayed conversion, but did not increase immediate conversion.Evidence A · P1
  450. MWITA-PRO-2026-009Passive subscription renewal frictions approximately doubled modeled seller revenue among a fixed set of initial subscribers.Evidence B · P1
  451. MWITA-PRO-2026-010A sudden reduction in a movie subscription's fixed fee increased downstream box-office revenue, especially for high-quality niche films.Evidence B · P1
  452. MWITA-PRO-2026-011A continuous-pricing demand model increased airline revenue in an operational field deployment, though the broader rollout estimate was smaller than the initial field result.Evidence B · P1
  453. MWITA-PRO-2026-012Selecting a non-discounted meal bundle increased purchases of products outside the bundle by causing shoppers to underestimate total basket size.Evidence A · P1
  454. MWITA-MIO-2026-001Across 15 Facebook advertising RCTs, commonly used observational estimators often failed to recover experiment-identified lift even with extensive demographic and behavioral controls; in half of the purchase studies every tested observational method missed percentage lift by at least a factor of three.Evidence A · P1
  455. MWITA-MIO-2026-002In 663 Facebook RCT benchmarks, neither double/debiased machine learning nor stratified propensity-score matching reliably recovered causal ad lift despite more than 5,000 user- and experiment-level features.Evidence A · P1
  456. MWITA-MIO-2026-003A randomized ghost-ad retargeting experiment found 17.2% more website visits and 10.5% more purchases while attaining comparable lift precision at at least an order of magnitude lower control-group media cost than intent-to-treat or PSA designs.Evidence A · P1
  457. MWITA-MIO-2026-004eBay randomized paid-search suppression and found no measurable short-run benefit from brand-keyword ads; for nonbrand terms, the total paid-search regime added an estimated 0.66% to sales with a 95% confidence interval spanning -0.42% to 1.74%, and average returns were negative because frequent users absorbed most spend without responding.Evidence A · P1
  458. MWITA-MIO-2026-005A Yahoo randomized display-ad experiment for a national retailer increased total purchases by about 5%; 93% of incremental sales occurred in physical stores and 78% came from people who never clicked an ad.Evidence A · P1
  459. MWITA-MIO-2026-006Twenty-five major-advertiser field experiments showed that even very large randomized ad tests often leave ROI too imprecise for confident budget decisions.Evidence A · P1
  460. MWITA-MIO-2026-007Free Yelp search advertising caused average gains of 19% in restaurant page views, 7%-14% in downstream purchase-intention actions and 5% in customer reviews during treatment, but tracked effects dissipated immediately after ads stopped.Evidence A · P1
  461. MWITA-MIO-2026-008Across 288 established consumer brands, estimated marginal TV-ad ROI was negative for more than 80% of brands and the overall observed advertising schedule had positive ROI for only about one-third.Evidence A · P1
  462. MWITA-MIO-2026-009Randomly switching on retargeting for visitors to a home-improvement retailer caused 14.6% more users to return within four weeks; 33% of first-week advertising's effect occurred on day one, and week-two ads worked better following nonzero week-one assignment.Evidence A · P1
  463. MWITA-MIO-2026-010A randomized European e-retailer retargeting test found about a 4.5% relative increase in purchase probability but no evidence that users receiving higher platform bids were more incrementally responsive; under the CPA contract, 46.4% of platform fees were tied to purchases estimated to have happened anyway.Evidence A · P1
  464. MWITA-MIO-2026-011Bing experiments showed brand ads without competitors increased focal-brand clicks only 1%-4%; competitors in positions 2-4 diverted 1%-5% when the focal ad remained, while competitors could take 18%-42% when a nonadvertising focal brand faced competition.Evidence A · P1
  465. MWITA-MIO-2026-012In three advertiser case studies benchmarked to randomized geo experiments, naive observational paid-search ROAS was roughly 15x, 8.4x and 2.9x the indexed experimental estimates; query-volume selection-bias correction reduced estimates to 3.0x, 1.9x and 0.8x, respectively.Evidence B · P1
  466. MWITA-PCR-2026-009Both traditional and social earned media predicted microlending ‘sales’; a traditional mention had the larger per-event effect, but social earned media had the larger elasticity after its much higher event frequency was incorporated and also preceded traditional coverage.Evidence B · P1
  467. MWITA-PCR-2026-010Corporate-misconduct events had no average US stock reaction until at least four high-reach outlets covered them; at that threshold mean three-day abnormal return was -1.12%, revealing exposure scale as a market-value boundary.Evidence B · P1
  468. MWITA-PCR-2026-011For 169 Chinese product-harm crises, firm-first disclosure was associated with a more negative three-day market reaction than third-party disclosure (-1.74% versus -1.09% CAR), the opposite of the trust benefit often found in consumer experiments.Evidence B · P1
  469. MWITA-PCR-2026-012After a five-month peanut-butter recall, post-crisis advertising accelerated first repurchase for the stronger Kraft brand but not the weaker Eta brand; observed recovery depended on pre-crisis loyalty, familiarity and brand strength rather than communication alone.Evidence B · P1
  470. MWITA-ELO-2026-001In the main randomized field experiment, adding the recipient's name to the subject line raised detected opens from 9.05% to 10.80%, sales leads from 0.39% to 0.51%, and reduced unsubscribes from 1.2% to 1.0%; related experiments across three firms also reported benefits.Evidence A · P1
  471. MWITA-ELO-2026-002Two randomized German field experiments did not reproduce a benefit from first-name subject-line personalization among students: detected opens were 3.2% lower and clicks similar, both nonsignificant; title-plus-surname personalization among politicians raised opens 11.8% significantly while its 14% click lift was nonsignificant.Evidence A · P1
  472. MWITA-ELO-2026-003Across three longitudinal field experiments, emails tailored with data applicants had actively volunteered outperformed observed-data personalization in aggregate open, click-to-open and conversion outcomes, but unsubscribe differences were not significant and ordinary observed-data personalization did not outperform the untailored control.Evidence A · P1
  473. MWITA-ELO-2026-004Reducing coupon-email frequency from daily to every other day for one week cut unsubscribe probability 59.2% but also reduced daily revenue 7.6%, purchase probability 4.8% and item quantity 10%; price per item was effectively unchanged.Evidence A · P1
  474. MWITA-ELO-2026-005Two incremental engagement emails over one month increased five-month subscriber retention likelihood 7.4%, reduced churn odds 26.3% and increased per-period service consumption 7.0%, yet modeled profit rose only for some subscription/use segments and fell for others because retained usage creates service cost.Evidence A · P1
  475. MWITA-ELO-2026-006In the stratified randomized beauty-retail experiment, email had no significant sales effect in any of six customer-value segments, despite observational models finding email responsiveness for some current-customer segments; direct mail worked for prospects and the email-plus-mail interaction was small but significant only for medium-value customers.Evidence A · P1
  476. MWITA-ELO-2026-007Across 70 randomized experiments on a ticket-resale platform, emailed discounts increased average expenditure by $3.03, or 37.2%, during the promotion window and by $1.55 the following week; about 90% of the promotion-window gain occurred without offer redemption.Evidence A · P1
  477. MWITA-ELO-2026-008Varying the subject line across weekly lifecycle emails produced small detected-open gains: 1.1 percentage points across weeks 2-3 among active/new learners and 0.3 points among all new learners; the fixed-effects meta-analysis estimated +0.7 points, while course progress was null.Evidence A · P1
  478. MWITA-ELO-2026-009Across three randomized field experiments, authority cues in sender names or subject lines sometimes improved early detected opens but did not improve clicks; the initially selected subject-line winner was not statistically significant when measured at 30 days, indicating a winner-timing/fade-out problem.Evidence A · P1
  479. MWITA-LRO-2026-001Crossing a Yelp display-rating threshold implies about 5%–9% higher revenue for independent Seattle restaurants; chain-affiliated restaurants show no rating response.Evidence A · P1
  480. MWITA-LRO-2026-007A 1% improvement in a hotel's aggregate online reputation score was associated with up to 0.89% higher average daily rate, 0.54% higher occupancy and 1.42% higher revenue per available room.Evidence B · P1
  481. MWITA-LRO-2026-008A 10% improvement in review rating was associated with 4.4% more hotel-booking proxy volume, while a 10% increase in rating variance was associated with 2.8% less.Evidence B · P1
  482. MWITA-LRO-2026-009Randomly soliciting a hotel review reduced the solicited customer's subsequent brand spending by $16, or 7%, on average; harm was concentrated among less-than-perfectly-satisfied, highly loyal customers.Evidence A · P1
  483. MWITA-LRO-2026-010After EU roaming charges ended, tourist-accessible high-rated Rome restaurants gained nearly 6% annual revenue, mid-rated restaurants about 4%, low-rated restaurants no revenue lift, and the exit probability of low-rated restaurants doubled.Evidence A · P1
  484. MWITA-LRO-2026-011In Expedia's randomized-ranking sample, click and booking probabilities decline materially with lower position; a modeled consumer-surplus ranking would outperform Expedia's ranking by $1.205 consumer surplus and $0.084 platform revenue per impression.Evidence A · P1
  485. MWITA-CUX-2026-001In randomized holdbacks on Stripe Checkout, surfacing at least one additional locally relevant payment method beyond cards increased completed purchase conversion by 7.4% and revenue by 12% on average; effects were highly heterogeneous by method and market.Evidence A · P1
  486. MWITA-CUX-2026-002In a 50/50 randomized holdback over more than 150,000 eligible Checkout sessions, displaying one eligible BNPL method increased conversion across order values and produced up to 14% more revenue; Stripe attributes more than two thirds of BNPL volume to incremental rather than displaced purchases.Evidence A · P1
  487. MWITA-CUX-2026-003Across 1.5 million subscription Checkout sessions with a randomized 1% holdback, presenting localized currency through Adaptive Pricing increased signup conversion by 4.7%, authorization by 1.9% and subscription lifetime value per session by 5.4%.Evidence A · P1
  488. MWITA-CUX-2026-004A large StubHub field experiment found that users shown mandatory fees only late in checkout were 13% more likely to purchase and spent about 21% more than users shown all-in prices upfront, while spending 5.42% more conditional on purchase; the revenue gain therefore coexisted with reduced price transparency.Evidence A · P1
  489. MWITA-CUX-2026-005In a randomized German cinema checkout experiment with more than 34,000 customers, shrouding a EUR 3 3D surcharge increased initiation of 3D purchases but also increased checkout dropout, leaving final 3D demand statistically unchanged across disclosure conditions.Evidence A · P1
  490. MWITA-CUX-2026-006Among 2,605 recruitment-site visitors, one infographic landing page reduced interest-form completion from 16.8% to 12.8% (OR 0.73, p=.019), while a second infographic was null; neither changed eventual study randomization.Evidence A · P1
  491. MWITA-CUX-2026-007Two randomized web experiments found that shorter commercial landing pages generated significantly more lead-form submissions than longer information-rich versions across locations and observation periods.Evidence A · P1
  492. MWITA-CUX-2026-008In a randomized eight-condition landing-page field experiment with 956 visitors, repeating the headline phrase verbatim on the call-to-action button increased click-through by more than 10 percentage points relative to introducing new wording.Evidence A · P1
  493. MWITA-CUX-2026-009In a randomized federal digital-form test, embedding instructions at the relevant fields raised submission from 33.0% to 36.2%, a 3.2 percentage-point estimate whose 95% confidence interval included zero.Evidence A · P1
  494. MWITA-CUX-2026-010In a preregistered-style online choice experiment with 5,889 UK participants, preselecting shipping or protection add-ons made consumers about 60%-70% more likely to select the more expensive option, whereas merely visually mimicking preselection without an actual default was null.Evidence A · P1
  495. MWITA-CUX-2026-011In a six-country randomized online experiment with 7,430 participants, tested dark-pattern interfaces increased inconsistent choices by 6-12 percentage points versus placebo interfaces, and a transparency remedy did not eliminate the distortion.Evidence A · P1
  496. MWITA-CUX-2026-012Among 75,914 new ecommerce visitors, requesting registration before browsing increased registrations by 58.08% but left 24-hour purchase statistically unchanged; among users induced to register, instrumental-variable estimates over 47 days indicated higher purchase probability, order count and revenue.Evidence A · P1
  497. MWITA-IMO-2026-005A conversational LLM agent raised municipal-web task success for blind and low-vision users from 80.0% with their screen reader to 95.8%, a 15.8-point gain.Evidence A · P1
  498. MWITA-IMO-2026-007Across all 21 Amazon regional markets, product-page accessibility improved over six months on the authors' index, but homepages were unchanged and some markets, including India and Canada, worsened.Evidence B · P1
  499. MWITA-IMO-2026-009InstaFetch consolidated dispersed product information and cut blind users' mean per-item shopping-task time by 61.9% versus a standard screen reader and 41.3% versus a state-of-the-art landmark aid.Evidence A · P1
  500. MWITA-IMO-2026-010QuickCue's accessible aspect-and-sentiment summaries improved usability and reduced workload when blind users compared restaurant reviews, but purchase or restaurant-choice effects were not observed.Evidence A · P1