Resolution types and when each is used

FeedbackRobot has several ways to resolve a piece of feedback. You control which are enabled.

Apology

A direct acknowledgement that something went wrong, referencing the specific problem.

Use for most negative feedback where the issue is clear and doesn’t need compensating — a slow wait, a mistake that’s already been corrected, a one-off lapse.

The value here is speed and consistency. An apology that arrives within the hour and names the actual problem does most of the work of service recovery on its own.

Thank you

A response to positive feedback that acknowledges what specifically the customer liked.

Use for positive reviews and high-scoring survey responses. It’s easy to dismiss this as low value — it isn’t. Customers who get a genuine response to positive feedback are meaningfully more likely to return and to review again. It also opens the door to asking for a public review from someone who’s already told you they’re happy.

Promo code

An apology backed by something concrete — a discount or offer on their next visit.

Use for problems significant enough that words alone won’t cover it, and where you want the customer back rather than just placated.

Two cautions:

• Set the threshold carefully. If every complaint produces a discount, you’ve taught people to complain.
• Make sure the code works. A recovery gesture that fails at the till is worse than no gesture.

AI conversation

Instead of a single message, the agent has an actual back-and-forth with the customer — asking what happened, digging into detail, and working toward a fix.

Use for situations where you don’t yet know what went wrong, or where the customer will likely reply. This produces by far the best information, because most people will explain the real problem if asked a follow-up question.

Don’t use for anything legally or medically sensitive. Configure those to escalate instead.

Tone and length

Each resolution type carries settings for tone (friendly, formal, and so on) and response length. Shorter is usually better for apologies — a long apology reads as defensive. Longer suits AI conversations where you’re genuinely trying to understand something.

→ Setting tone and response style

Choosing what to enable

A sensible starting configuration:

Feedback	Resolution
Positive review or high survey score	Thank you
Mild complaint, clear cause	Apology
Serious complaint, customer worth recovering	Promo code
Unclear complaint	AI conversation
Safety, legal, medical	Escalate to a human

Start conservative. It’s easier to widen the agent’s latitude once you’ve read a few weeks of real replies than to walk back an automated offer that went to the wrong person.