Re:amaze AI quality assurance: what its reports and AI tools check, and what they don't

Riellvriany Indriawan
Written by

Riellvriany Indriawan

Katelin Teen
Reviewed by

Katelin Teen

Last edited November 22, 2026

Expert Verified
Hand-drawn illustration of a support lead at a laptop and a colleague pointing at a monitor showing three replies with teal check marks and a robot with a checklist, with the Re:amaze logo on a teal background
TL;DR

Re:amaze has no AI quality assurance scorecard: I found nothing in its help center, pricing page or product news that scores a conversation against QA criteria, and the closest it gets is the recent-replies list in the Individual Staff report on Plus ($69 per team member a month). What it gives a support lead is reports on speed and how often one reply solved the issue (every plan), staff reports and satisfaction surveys (also Plus), and a set of tools for reviewing its own AI Agent's answers. None of them checks whether a reply was right for your policy or tagged the right way, and the AI report's "Success Rate" counts successful API requests, not right answers. To check that, a lead can score a weekly sample by hand, or an AI teammate like eesel can review recently archived conversations on a schedule against rules you write and leave its reasons in an internal note. A coding agent such as Claude Code can set that review up through the eesel CLI, though a person still copies the API token from Re:amaze to connect it.

Does Re:amaze have AI quality assurance?

Not under that name, and not as a scorecard. I work on eesel's support team, so when I read a helpdesk's QA features I'm asking one thing: would this catch a wrong reply before a customer does? In Re:amaze, the honest answer is "only after a customer complains". (For the rest of the product, see my Re:amaze review.)

Here is what Re:amaze does offer for quality assurance, from its reporting overview, AI tools guide and pricing page, checked 22 November 2026:

Re:amaze featurePlanWhat it tells a QA leadWhat it doesn't
Response Time, First Response Resolve Rate, Average Thread SizeBasicHow fast replies went out and how often one reply settled itWhether the reply was right
Appreciations, Tags and Satisfaction Ratings reportsPro, per the reporting overviewWhich replies customers liked, which tags get used, what customers ratedWhich tag should have been used
Staff Summary and Individual Staff reports, satisfaction surveysPlusEach agent's responses, response times, ratings and recent repliesWhether those replies followed your process
AI Agent sources, corrections, testing and Learned AnswersWith the AI AgentWhere an AI answer came from, and a place to fix itHow human agents are doing
Ask About This ConversationAI toolsAn answer to a question you type about one threadAnything across many conversations
Hand-drawn chart titled What Re:amaze tells you about quality, with How fast replies went out, How often one reply solved it, What customers rated after the fact and Which AI answers you corrected checked under Measures, and Was the answer right for our policy, Was it tagged and routed right and Did the rep follow the process crossed under Never checks, noted as Staff and satisfaction reports on the Plus plan
Re:amaze measures the outcome. The content of the reply is left to you.

What do Re:amaze's reports say about quality?

They say a lot about speed and a little about outcome. Every plan gets the Response Time report (the first response time glossary entry explains the metric most teams start with) and the First Response Resolve Rate, which Re:amaze describes as "a quality metric that reflects the effectiveness of your team's responses": how often one reply settled the customer's issue, close to what most teams call first contact resolution. The Average Thread Size report is its partner, and Re:amaze's own guide says a high number "can indicate that an issue is difficult to resolve or that agents are not providing sufficiently complete responses."

On Plus, the Individual Staff report breaks this down per agent: responses, average response time, appreciations, conversations resolved or archived, average thread size and, if you send surveys, the agent's average rating. You can compare an agent with the team average or one department. At the bottom sits "a running list of their recent replies", which Re:amaze says managers can use "to monitor support quality at the individual level."

That last line is the closest Re:amaze gets to QA. It gives you a reading list; someone still opens each reply and judges it.

Two things to know before you trust the numbers:

  • Conversations are dated by when they started. Re:amaze attributes conversation metrics "to the time the conversation was created", so a conversation that opened on 9 December and was reopened and archived on 11 December counts once, on 9 December. Last week's numbers can move after the week ends.
  • Responses are counted per message. Re:amaze says the Response count "is a better indicator of activity" than the Resolved and Archived counts.

What do Re:amaze's satisfaction surveys add?

The customer's own verdict, late. Automated satisfaction surveys are on Plus: under Settings, then Email Channels, turn on Satisfaction Surveys, pick a delay in hours and which conversation states send a survey (choosing Resolved sends one only after a conversation is resolved). The survey can use smiley faces, descriptive labels or 1 to 5 stars, and the Satisfaction Ratings report shows the average, the spread from Abysmal to Great, and the latest comments.

There's a plan wrinkle. The pricing page and the survey guide put satisfaction ratings on Plus, but the reporting overview lists the Satisfaction Ratings report under Pro. If you're on Pro, ask Re:amaze before you plan around it.

A rating tells you how the customer felt about the outcome. It doesn't tell you which reply caused it, and it only comes from the customers who answer. A low score tells you where to start reading. My notes on AI CSAT go further on reading CSAT next to QA.

How do you check the Re:amaze AI Agent's answers?

This is where Re:amaze has done the most work. In June 2026 it shipped a set of tools for reviewing what its AI Agent says:

  • Sources on every reply. A source count appears on each AI response; click it to see which knowledge sources it used, "a quick way to spot if the AI leaned on something outdated or pulled from the wrong place."
  • Improve Response. Submit a correction on any AI reply, and it's stored for the AI to use next time. Re:amaze says corrections are kept in one place where you can edit or delete them, and the AI settings have an AI Agent Corrections tab.
  • Test AI Agent. Ask it the questions customers ask and read the full answers before a customer does. Re:amaze suggests coming back to it whenever you change an article, a template or a learned answer.
  • Learned Answers. When the AI Agent escalates and a person finishes the conversation, Re:amaze drafts an answer from what happened and queues it for review. Nothing is added until you save it, and unsaved suggestions expire after 14 days.
Re:amaze AI Chatbots settings with tabs for Custom AI Context, Learned Answers, AI Agent Corrections, Test AI Agent and AI Pre-Filter, showing a Learned Knowledge suggestion for review about free expedited shipping during Black Friday, with Save and Delete buttons and View Conversation and Edit links
Corrections, testing and learned answers sit side by side in the AI settings, as shown in Re:amaze's Learned Answers guide.

These are good tools for keeping the AI Agent honest, and I'd use every one. Note what they have in common, though: a person finds the bad answer first. Nothing scores the AI Agent's conversations for you.

Watch the AI reports too. The AI Productivity Report has a Success Rate, which "measures the number of successful AI API requests", and a Content Score, a setup checklist (+20% for creating 20 FAQ articles, +20% for setting up Custom AI Context). A 100% Success Rate means the AI ran every time, not that it was right every time.

What can't Re:amaze check?

The three things most QA programs exist for:

  1. Was the answer right for our policy? No report or AI tool compares a reply with your refund, shipping or warranty rules. Ask About This Conversation will answer a question you type about one thread, but that's one conversation at a time, by hand.
  2. Was it tagged and routed right? AI Pre-Filter and Workflows apply tags to conversations, and the Tags report counts which tags are applied most often, but nothing checks them afterwards. Neither tells you a refund request went out with the shipping tag, and if your reports run on tags, that miss spreads into every chart.
  3. Did the agent follow the process? Asked for the order number, offered the return label, escalated a chargeback threat: none of it is scored.

If you look at a dedicated QA tool instead, check that it connects to Re:amaze before you buy.

How much does Re:amaze cost for QA?

The QA-relevant reports and surveys come with the plan, so the cost is the step to Plus. Prices from the pricing page, checked 22 November 2026:

PlanPer team member a monthBilled annuallyWhat it adds for QA
Basic$29$26.10Response Time, First Response Resolve Rate, Average Thread Size; AI Agent with 5 resolutions per user a month
Pro$49$44.10Advanced Reporting (Tags, Appreciations, Workflow); 10 AI Agent resolutions per user
Plus$69$62.10Staff Performance Reporting, Customer Satisfaction Ratings; 20 AI Agent resolutions per user
Starter$59 flat a monthNot listedBasic's features for unlimited team members, up to 500 responded conversations a month

Extra AI Agent resolutions are $0.85 each. For a team of 10 paying monthly, Pro is $490 and Plus is $690, so per-agent reports and surveys arrive with a $200 a month step up. My Re:amaze pricing breakdown and Re:amaze cost guide have other team sizes.

Re:amaze pricing page set to annual billing, showing Basic at $26.10, Pro at $44.10 and Plus at $62.10 per team member, with Staff Performance Reporting and Customer Satisfaction Ratings listed under Plus
Staff Performance Reporting and satisfaction ratings start on Plus, as shown on Re:amaze's pricing page.

How do you cover what Re:amaze leaves out?

Pick what you want to check and see what covers it:

What do you want to check on a Re:amaze conversation?
Re:amaze covers it: Response Time and First Response Resolve Rate
Both reports are on every plan. First Response Resolve Rate counts how often one reply settled the issue, which Re:amaze calls a quality metric.
PlanBasic and up
WhereReports
Counts byConversation date
Read Average Thread Size next to it: long threads can point at answers that weren't complete.
Re:amaze covers it on Plus: satisfaction surveys
An email survey goes out a set number of hours after a conversation reaches the state you pick, and the Satisfaction Ratings report collects the scores and comments.
PlanPlus, per the pricing page and survey guide
ScaleAbysmal to Great
AlsoAppreciations report on Pro
A rating comes late and only from customers who answer, so it can't tell you which reply went wrong.
Not in Re:amaze: a reviewer who knows the policy
No Re:amaze report or AI tool checks a reply against your refund, shipping or warranty rules.
By handA lead reads a weekly sample
With AIA reviewer that reads your help center and the thread
Write firstThe rules, one line each
This is the check most QA programs exist for, so start here.
Not in Re:amaze: the Tags report only counts
The Tags report shows which tags are applied most often. Nothing checks that a refund request got the refund tag.
By handCheck tags while reading the sample
With AIA reviewer that reads tags with the thread
Write firstWhich tag each topic gets
If your reports run on tags, a wrong tag is a QA miss even when the reply was perfect.
Partly: Re:amaze's AI Agent review tools
Each AI reply shows the sources it drew from, Improve Response stores a correction, Test AI Agent lets you ask it questions, and Learned Answers queues answers from escalated conversations for review.
CorrectionsAI Agent Corrections tab
Before go-liveTest AI Agent
Watch forSuccess Rate counts API requests, not right answers
Someone still has to open the replies: none of these scores a conversation for you.

Pick a check to see what covers it in Re:amaze, and what covers it if nothing does.

There are two ways to cover the rest.

Score a sample by hand

The cheapest route needs nothing new, and it's where most guides to support QA with AI start. Each week a lead takes ten or twenty recently archived conversations (the longest threads and the lowest ratings first, if you have them), reads each, and scores it against a short rubric in a spreadsheet: right answer for the policy, right tag, process followed. On Plus, the recent replies list in each Individual Staff report is a ready-made reading list. Agree what each score means first so two reviewers score the same conversation the same way: support QA calibration helps, and so do these QA feedback examples.

The limit is the obvious one: a person reads a sample, and the results live outside Re:amaze.

Have an AI teammate review against your rules

The second route is an AI agent that runs the review on its own and leaves the result in Re:amaze. This is what I'd do with eesel. eesel's Re:amaze integration runs on a schedule, as often as every ten minutes, or hourly, daily or weekly, and can list archived conversations and narrow them by tag. For each one it reads the newest 40 messages with their authors, internal notes included, plus the conversation's tags, status and assignee. It learns from your response templates and past conversations ahead of time and reads your help center articles live, so a policy you've published is something it has read.

Set a daily schedule and, on each run, it takes the most recently changed archived conversations and:

  1. Reads the thread with its tags.
  2. Checks it against the rules you wrote in plain words ("refunds after 30 days need a manager's approval", "billing questions get the billing tag", "never promise a ship date").
  3. Leaves its verdict and reasons as an internal note.
  4. Adds a tag such as qa-review to the ones a lead should read, so a lead can filter for them.
A Re:amaze conversation from a first-time skier asking what equipment to hire, with an eesel internal note underneath holding a drafted reply that lists skis, boots, helmet and poles
eesel writes into Re:amaze as an internal note, which only your team sees.

You write the rules, so tagging, process and policy checks are just more lines, changed whenever your policies change. Know the limits first: it doesn't see attachments or custom fields, and on a thread longer than 40 messages it reads only the newest 40. It runs on a schedule rather than the moment a conversation closes, because Re:amaze has no event triggers in eesel. The same eesel agent can draft replies as internal notes too, if you want it to.

Daily checks like this already run for eesel's customers. Cactus Club Cafe runs a second eesel agent on its Zendesk helpdesk that audits the daily finance extract every morning: it checks each store and either solves the ticket as balanced or leaves it open with the imbalance named. Between 8 August and 8 September 2026 it ran about 60 times. A QA review is the same shape: a check against written rules, with the result left on the conversation.

Is Re:amaze enough for QA?

For speed and outcome, yes. First Response Resolve Rate and Average Thread Size are honest signals that most small teams under-use, and on Plus the per-agent report with ratings and recent replies gives a lead a sensible weekly reading list. The AI Agent tools are better than most: sources on every reply, stored corrections and a test environment are exactly what you need to keep an AI agent from drifting.

My take: start with the reports you already have. Watch First Response Resolve Rate, and Average Thread Size in each Individual Staff report, read the longest threads every week, and use Test AI Agent whenever you change an article. Don't mistake the AI report's Success Rate for accuracy. And if your team's QA question is "was it right?", not "was it fast?", Re:amaze leaves that to you: a hand-scored sample if your volume is small, an AI reviewer if it isn't. Other helpdesks' answers to the same question are in my posts on Gorgias AI quality assurance, Help Scout AI quality assurance and Front AI quality assurance, and my roundup of AI tools for support QA compares dedicated options.

Try eesel as your Re:amaze QA reviewer

If the check you're missing is "was this reply right for our policy?", eesel is the reviewer I'd add to Re:amaze. It connects with an API token, reads your help center and past conversations, and on a daily schedule checks each recently archived conversation against the rules you write, leaving its reasons in an internal note and tagging the ones worth a lead's time. Start by asking it to review your most recent archived conversations in the dashboard chat, read what it flags, and only then turn the schedule on.

eesel's schedule editor for a Re:amaze automation, with Minutes selected, Every 10 minutes in UTC, and an Instructions box describing which Re:amaze conversations to work and what to do with them
A review schedule is a timer plus instructions in plain words. Switch it to Daily for QA.

Try eesel free, or book a demo to see it review your own Re:amaze conversations.

Frequently Asked Questions

Does Re:amaze have AI quality assurance?
Not as a scorecard. I found no feature in Re:amaze's help center, pricing page or product news that scores conversations against QA criteria. It has reports on speed, one-reply resolution and customer ratings, plus tools for reviewing its AI Agent's answers. AI support quality assurance covers what a full program checks.
Which Re:amaze plan has staff performance reports?
Plus, at $69 per team member a month or $62.10 billed annually, per the pricing page on 22 November 2026. Re:amaze's reporting overview lists the Staff Summary and Individual Staff reports under Plus. Re:amaze pricing has the full breakdown.
Does Re:amaze send customer satisfaction surveys?
Yes, on Plus. You turn surveys on under Settings, then Email Channels, pick a delay in hours and which conversation states send one, and choose smiley faces, descriptive labels or 1 to 5 stars. Re:amaze's reporting overview lists the Satisfaction Ratings report under Pro, so check with Re:amaze if you're on Pro. CSAT explains what the score means.
What does Success Rate mean in Re:amaze's AI reports?
It measures the number of successful AI API requests, per Re:amaze's AI report guide, so it tells you the AI ran, not that its answers were right. Content Score is a setup checklist, such as +20% for 20 FAQ articles. My Re:amaze review covers the rest of the product.
How do I review the Re:amaze AI Agent's answers?
Each AI reply shows a count of the sources it drew from, Improve Response stores a correction for next time, Test AI Agent lets you ask it questions before customers do, and Learned Answers queues suggested answers from escalated conversations for you to save or delete. Re:amaze AI draft replies covers the AI Agent in more depth.
Can Re:amaze check whether conversations were tagged correctly?
Not on its own. The Tags report counts which tags are applied most often, but nothing in Re:amaze checks that a conversation got the right tag. Re:amaze AI ticket tagging and routing covers how tags get applied in the first place.
Can an AI agent review Re:amaze conversations against my own policies?
Yes. eesel connects to Re:amaze, runs on a schedule as often as every ten minutes, or hourly, daily or weekly, reads recently archived conversations with their tags and internal notes, checks each against rules you write, leaves its reasons in an internal note and adds a review tag. Re:amaze automation covers what else it can take on.

Share this article