# Intercom Fin Resolution Rate: How to Measure It

*How-to — 2026-10-18 — by Mahmoud Zalt*

Where Fin's resolution rate lives, why automation rate is usually the lower number, and how assumed resolutions change what the figure means.

**Short answer.** **Resolution rate and automation rate answer different questions, and the gap between them is the number most people are actually looking for.** Both sit on the Support Performance dashboard under Fin AI Agent, then Analyze, then Performance. Resolution rate is measured against conversations Fin was involved in. Automation rate is measured against every new conversation, which is why it is usually the smaller figure.

**About this guide.** This guide is based on publicly available information and Intercom's own documentation, checked on 17 August 2026. Sistava is not affiliated with Intercom. Interfaces and pricing change often, so confirm the current steps in their documentation before you rely on them.

Someone will ask you what percentage of support Fin is handling. It is a fair question with an annoying answer, because Fin reports at least four numbers that could all honestly be described that way, and they differ by a lot.

This guide covers where each number lives, what its denominator is, what counts as a resolution in the first place, and how to sanity check a figure that looks too good.

None of this is Intercom hiding anything. Their documentation defines the metrics openly. The confusion comes from people quoting one number in a meeting where a different one was the honest answer.

## Where do you find Fin's resolution rate?

Two places, for two different jobs. The Support Performance dashboard is at Fin AI Agent, then Analyze, then Performance, and Intercom describes it as bringing four core metrics together: automation rate, resolution rate, involvement rate and CX Score.

For anything deeper, use the Fin AI Agent report template in the Reports section. That is where the full metric list lives, along with date filters, channel breakdowns and a content performance table.

## What is the difference between resolution rate and automation rate?

The denominator. Intercom defines automation rate as the percentage of conversations resolved by Fin out of all new conversations. Resolution rate is measured against conversations Fin was involved in, which is a smaller pool.

So a workspace where Fin only sees a third of incoming conversations can post a high resolution rate and a modest automation rate at the same time, and both are true. If your board wants to know how much of the queue is being handled, automation rate is the honest number.

| Metric | What it measures | Use it to answer |
|---|---|---|
| Automation rate | Conversations resolved by Fin out of all new conversations | How much of the total queue is Fin taking off us? |
| Resolution rate | Conversations Fin resolved, out of conversations Fin was involved in | When Fin gets a shot, how often does it land? |
| Involvement rate | How often Fin participates in conversations | Are we even letting Fin see the traffic? |
| CX Score | The share of positive ratings out of all ratings received, for conversations handled by Fin alone without a human | Do customers like the answers they got? |

**A low automation rate is often a routing problem.** If resolution rate is healthy and automation rate is not, Fin is answering well but rarely getting the chance. Look at your involvement rate and your deployment audience rules before you touch a single article.

## What actually counts as a resolution?

Intercom's outcomes documentation defines a resolution as no further help being requested after Fin's last answer. That definition splits into two kinds, and the split matters more than most teams realise.

- **Confirmed resolution. **The customer actively indicates they are satisfied. This is the strong signal.
- **Assumed resolution. **The customer leaves without asking for more help. Fin counts it as resolved because nothing contradicted it.

Assumed resolutions are a reasonable default, and every vendor in this category needs some version of them, because most people simply close the tab when they are done. But a customer who gave up and emailed you separately looks identical to a customer who was helped.

The good news is Intercom reports these separately. The Fin AI Agent report template includes both assumed resolution rate and confirmed resolution rate as distinct metrics. Pull both. The ratio between them tells you how much of your headline figure is evidence and how much is silence.

There are other outcomes on that page worth knowing, because they explain conversations that appear in neither column. Intercom lists procedure handoff, qualification and disqualification for sales cases, escalation, abandoned where a customer stops replying after a clarifying question, spam, and pending where an outbound message never got a reply.

## Why did my resolution rate change without me doing anything?

Intercom changed how these metrics are calculated, and published a page explaining it. Conversations where Fin was active but never got the chance to answer are no longer counted as Fin being involved. Their examples include escalations, triggered workflows, usage limits, and customers who never asked a question.

The mechanical effect is that involvement rate goes down and resolution rate goes up, because those conversations leave the denominator. Automation rate is unaffected. Intercom is explicit that the underlying number of resolutions did not change, only the reporting.

Their page gives dates for the switchover, with new workspaces created on or after 25 June 2026 using the updated definitions and existing workspaces moving over between 1 and 8 July 2026. If you have a chart that steps up around then, that is why, and you should not report it as an improvement.

**Check the definition before you compare two periods.** Any metric definition change breaks year on year comparisons at the moment it lands. Note the change on your dashboard, keep the old figures labelled as the old calculation, and compare like with like.

## What is a good resolution rate?

There is no single right answer, and be careful with any benchmark you see quoted without a denominator attached. A number measured against the conversations an agent was involved in is not comparable to one measured against a whole queue.

Independent practitioner writing gives a sense of the spread rather than a target. One analysis of Fin resolution rates reports a range from roughly 25 percent to over 80 percent across Intercom customers, and cites an Intercom support engineer describing 30 to 50 percent as a reasonable starting point. Treat those as context from a third party, not as Intercom's official guidance.

The same article describes a support lead whose dashboard showed a high resolution figure while their own review of conversations put the real rate far lower. Whether or not those exact numbers generalise, the lesson does: sample the transcripts yourself.

## How do you sanity check the number?

Read conversations. It is the only check that cannot be gamed by a definition, and an hour of it is worth more than a month of dashboard watching.

### A monthly resolution rate review

1. **Pull automation rate, not just resolution rate** — Report both, always, with their denominators named. This single habit prevents most of the disagreements about whether the tool is working.
2. **Split assumed from confirmed** — Both are available as separate metrics in the Fin AI Agent report template. A resolution rate made almost entirely of assumed resolutions deserves scrutiny.
3. **Read twenty transcripts at random** — Pick from conversations marked resolved. Count how many you would personally have considered resolved. That percentage is your correction factor.
4. **Check CX Score alongside it** — Intercom defines CX Score as the share of positive ratings among all ratings, for conversations Fin handled without a human. A high resolution rate with a weak CX Score is a warning.
5. **Break down by channel** — The report supports channel breakdowns. Chat and email behave differently, and one weak channel can hide inside a healthy average.
6. **Use the content performance table** — It can be sorted by involved conversations or by resolved conversations, which points straight at the articles doing the work and the ones nobody ever needed.

## A note on what the number is for

Resolution rate exists partly because per outcome pricing needs a definition of success. That is a reasonable trade and Intercom is transparent about the mechanics, but it does mean the metric is doing two jobs at once: telling you how well the agent works, and determining what you owe.

We took a different route at Sistava. Rather than metering resolutions, you hire an AI Employee on a plan and tell it in plain English what to handle, so the question of whether a given conversation counted as resolved never has to be settled by a definition. You still measure quality, you just do not have to negotiate with the meter while you do it.

In fairness, Fin's reporting is more detailed than most tools in this space, and the fact you can separate assumed from confirmed resolutions at all is a point in its favour. Plenty of vendors give you one number and no way to interrogate it.

## FAQ

### Where do I see Fin's resolution rate?

On the Support Performance dashboard at Fin AI Agent, then Analyze, then Performance. For a fuller breakdown, use the Fin AI Agent report template in the Reports section.

### What is the difference between resolution rate and automation rate in Intercom?

Automation rate is the percentage of conversations Fin resolved out of all new conversations. Resolution rate is measured against the conversations Fin was involved in. Automation rate is usually the lower and more conservative figure.

### What counts as a resolution for Fin?

Intercom's outcomes documentation defines a resolution as no further help being requested after Fin's last answer. That covers confirmed resolutions, where the customer indicates satisfaction, and assumed resolutions, where the customer simply leaves.

### Can I see confirmed and assumed resolutions separately?

Yes. Intercom's reporting documentation lists confirmed resolution rate and assumed resolution rate as separate metrics in the Fin AI Agent report template.

### Why did my Fin resolution rate jump in mid 2026?

Intercom changed how the metrics are calculated. Conversations where Fin was active but never got a chance to answer are no longer counted as Fin being involved, which raises resolution rate and lowers involvement rate. Their page states new workspaces from 25 June 2026 and existing workspaces between 1 and 8 July 2026.

### Does the metrics change mean Fin resolved more conversations?

No. Intercom is explicit that the actual number of resolutions and automation outcomes stayed the same, and that only the calculation and reporting changed.

### What is a good Fin resolution rate?

There is no official target, and any benchmark is meaningless without its denominator. Independent practitioner analysis reports a wide spread across customers and cites 30 to 50 percent as a reasonable starting point, but the reliable check is reading your own transcripts.

## Sources

- [Intercom Help, ](https://www.intercom.com/help/en/articles/8205718-fin-ai-agent-outcomes)Fin AI Agent outcomes
- [Intercom Help, ](https://www.intercom.com/help/en/articles/7837533-fin-ai-agent-reporting)Fin AI Agent reporting
- [Intercom Help, ](https://www.intercom.com/help/en/articles/11390083-monitor-fin-s-performance-with-clarity-and-confidence)Monitor Fin's performance with clarity and confidence
- [Intercom Help, ](https://www.intercom.com/help/en/articles/15599377-update-to-fin-performance-metrics)Update to Fin performance metrics
- [Pageloop, ](https://www.pageloop.ai/blog/how-to-improve-intercom-fin-resolution-rate)How to improve Intercom Fin resolution rate

**Tags:** intercom-fin, resolution-rate, support-metrics, ai-customer-support, reporting