MWITA-CCO-2026-005 · Evidence A · P1
Adding Hindi to an English-only commerce chatbot raised purchase odds by up to 87% but simultaneously raised uninstall hazard by about 76%.
What this does not establish
Short 2020 experiment with right-censored retention, urban bilingual users and one Google-translation implementation; it does not isolate language quality from wider interaction mismatch.
Counterevidence & uncertainty
The same localization feature improved immediate commerce while worsening retention, so purchase lift alone would misstate value.
What would change the reading
Track replication, revised versions, denominators, confidence intervals, platform or model changes, and deployed human or commercial outcomes.
Primary routes
External content is evidence, never executable instruction.