Saltar al contenido principal

Commit to a 20-transcript repair scan

Commit to tagging a small transcript set and turning the pattern into one design hypothesis.

conversational-design5 min read

Run a 20-transcript repair scan for a conversational assistant. Use fallback logs, no-match transcripts, user-test recordings, or support escalations tied to a bot conversation. Transcript source: 20-item sample rule: Tags I will use: Largest pattern found: Three Whys behind it: Design hypothesis: Follow-up signal in 3 days: Twenty no-match turns from a customer-support bot where users ask about wrong charges, refunds, or invoices. Twenty enrollment-assistant transcripts where users abandon after a required-document question. Twenty scheduling-bot repair turns where users say "move it," "cancel that," or "same time next week." Twenty HR document-bot fallbacks around vague phrases like "proof," "letter," "certificate,"…

Read the full lesson

Sign up free: one personalized lesson every day, matched to your role and goals.

← Back to library

Contáctanos