Skip to main content
Velor
All posts
PerspectiveNovember 2026 · 5 min read

Resolved or just gone quiet? Use re-contact rate to check any AI support number

By Caden Boatright


A conversation that ends without a human is not the same as a conversation that was solved. The customer may have got their answer. They may also have given up, closed the tab and written to you again an hour later through a different channel. A headline figure for deflection or resolution cannot tell those two apart. Re-contact can.

What re-contact rate measures

Re-contact rate is the share of customers the AI closed out who come back about the same issue within a window you choose, for example seven days. Reopen rate is the narrower version, for tickets that are formally reopened. Both are honest because the customer produces them, not a model grading its own answers.

How to check it with data you already have

  1. Pull the conversations the AI closed without a person in a past period.
  2. Match each customer, by email or account ID, to any new conversation or ticket they opened within your window, on any channel.
  3. Read a sample of the matches by hand. Some are new questions. The ones that repeat the original question are the misses.
  4. Split the result by topic, not just overall. A refund question and a password reset do not fail in the same way, and an average hides it.

Where it misleads

Re-contact undercounts when a customer gives up and never returns, so it is a floor on the problem, not a measure of all of it. It overcounts when a customer returns with a genuinely new question. Treat it as one check next to customer satisfaction scores and a hand-read sample, not as a replacement for them.

Questions to ask any vendor

  • Can I see re-contact for the conversations you counted as resolved?
  • Does a customer who comes back count as one resolution or two?
  • Is the figure split by topic, and can I read the conversations behind it?

Where Velor fits

Before you have production data, the question is whether the agent answers your tickets correctly at all. Velor's free eval answers that. You send 300 to 500 of your resolved tickets, with the customer's full message and your agent's reply, and Velor replays each one against your own help center. You get the resolution rate with a 95% confidence interval and every miss listed. We sign an NDA first, and names and emails can be stripped before you send anything.

The eval is a replay of past tickets, so it does not measure re-contact. That check belongs on live conversations. The Velor dashboard shows every conversation, handoff and resolution, with analytics and CSAT, so you can read what sits behind any figure.

We do not quote a resolution rate here. We have no published customer results, and the number depends on your tickets.

To start, request the free eval. For related reading, see who decides what counts as resolved, why resolution rate beats deflection and how to measure AI support.

Numbers in this post that cite no source are rules of thumb, not Velor measurements. Velor has no customer deployment to measure yet.

Velor

Want to see how this applies to your support queue?

Send us your recent tickets — at least 300, ideally 500 — and get a resolution rate, before you commit to anything.

14-day free trial · no credit card