Auto-resolve UX: act first, surface an undo, do not ask
Designing the human-AI handoff for a triage queue where an LLM-scored fuzzy match could attribute incoming messages to existing records
When the system has high enough confidence to act (≥0.85 fuzzy match + classifier verdict), the worst UX is showing the user a Resolve button with the candidate pre-selected — that is still asking them to do labor while pretending not to. The right shape: do the action, log it to a TTL-bounded undo journal (24h), and surface it under an auto-resolved · undo band lower in the page with a one-click revert that removes the link AND moves messages back. The Resolve button only appears for things the system was NOT confident about. This flips the framing from look how smart I was, please confirm to I did this, tell me if I was wrong — fewer clicks, much higher signal-to-noise, and an undo journal is easier to reason about than a permissions-and-prompts dance.
Sketch the action-then-undo flow first when designing any AI-augmented review UI; default to it unless the action is genuinely irreversible. A _autoresolved.jsonl file with at/person_id/identifier/moved_event_ids is enough scaffolding — no need to design an undo table upfront.