Feedback surveys

A feedback survey goes to the customer when a ticket is resolved, and the response appears on the ticket and on the customer's profile. It is at Settings → Ticketing → Feedback survey.

There is one survey per workspace. Reading it is open; editing needs admin or owner. Everything on the screen saves together, with Save survey.

A live Customer preview sits beside the form, so you can see the survey as the customer will rather than imagining it.

Delivery

Send a survey on resolution — the master switch. When off, no survey is sent.

Delivery channels — Email, SMS and WhatsApp, at least one required. Each carries a status: Available, or Outbound not connected. Your choice is stored and honoured automatically once that channel's outbound is connected, and nothing is sent on it until then. In practice, email is the channel that sends today; ticking the other two is a decision recorded in advance, not a delivery.

Send delay — minutes to wait after resolution, from 0 to 43200 (thirty days). Zero sends immediately. A short delay is usually worth it: it puts a little distance between "we closed your ticket" and "how did we do", and it gives a customer who is about to reply the chance to reply first.

Reminders — up to five follow-ups if the survey goes unanswered, with a Reminder interval of 1, 2, 3, 4 or 7 days. Zero reminders disables the interval control.

The survey itself

Survey type is one of four, and choosing one prefills the scale and the question:

• CSAT: Customer Satisfaction — 1 to 5, "How would you rate the support you received?"
• NPS: Net Promoter Score — 0 to 10, "How likely are you to recommend us to a friend or colleague?"
• CES: Customer Effort Score — 1 to 7, "How easy was it to get your issue resolved?"
• Custom — 1 to 5, "How would you rate your experience?"

Changing type only overwrites your question if it is still one of the presets, so a question you wrote yourself survives switching between types.

Question is up to 500 characters, and Scale from / Scale to are numbers from 0 to 10, with the maximum necessarily above the minimum. You can build any of the standard instruments here, but there is a real cost to inventing your own: a bespoke 1-to-6 scale is not comparable with anybody's benchmark, including your own from last year.

Copy

Follow-up prompt — up to 500 characters, shown under the rating as an optional comment box. Leave it blank to omit the box entirely. This is where nearly all the useful information comes from; a score tells you there is a problem and a sentence tells you what it is.

Thank-you message — up to 1000 characters, shown after submission.

Additional questions

Extra questions below the rating, each with a prompt, an answer type — Rating (1–5), Scale (1–10), Free text or Multiple choice — and a Required flag. Multiple choice takes its options one per line. Questions can be reordered and removed.

Restraint pays here. Every extra question costs response rate, and the response rate is what makes the score mean anything. One rating and one open box is a survey people finish.

Reading the summary

The right-hand column summarises what will actually happen: which channels it sends on, which are waiting on an outbound connection, how many reminders and how often, and the delay. When the survey is off it says so in one line, which is a quicker check than reading the switch.

This is a helpdesk feature rather than an AI one — no model is involved in sending it or in reading the answers.