Topic
twi-voice-agents
The Menu That Sent Adwoa Back Three Times, and What It Almost Cost Her
A main-menu loop can hide a failed Twi first contact. Learn what to track before a dashboard calls the interaction normal.
A caller switches to Twi. Their real question goes unanswered.
A clear English bot reply can still leave a Twi caller unanswered. Learn what support teams should test and record.
What Happens When a Caller Switches Between Twi and English?
Test Twi-English voice agents where it counts: code-switching, caller intent, consent, and the next step after the call.
The English Disclosure Ama Had, and Why It Failed in Twi
When an AI voice sounds vulnerable, disclosure can fade from view. Learn how to keep identity, consent, and opt-out clear throughout every call.
Arthur Stephenson’s investigation. Incompatible units had cost NASA a Mars orbiter.
Before launching a large calling programme, test consent, opt-outs, call outcomes and billing against what actually happened on the line.
What Happens When a Voice Agent Ignores a Customer’s Opt-Out?
Absa’s AI scale puts a sharper question on the table: can customer-facing agents prove consent, opt-outs and every call outcome?
Kwame’s Third Call, the One That Finally Found His Mother’s Money
A disbursement call in Twi exposes the gap between AI "decision-making" and real conversation. Why native speech, not model selection, decides whether public-service calls work.