Voice of customer · Intercom sentiment analysis
Intercom sentiment analysis that shows which product problems drive negative conversations
Product type
Ready to analyze
has signals waiting across usage, tickets and reviews. Surface the insights to see why users churn and what to build next.
Surfacing insights
AnalyzingHeadline insight
← friction step ยท biggest drop-off
Live and interactive, nothing is uploaded
Every insight traced to its source signals · aggregate & consented · no PII exposure
Short answer
Intercom has native sentiment analysis through CX Score, which rates each eligible conversation from 1 (negative) to 5 (positive) and gives the reasons. It needs the Pro add-on, billed at $0.20 per conversation with a 500 conversation monthly minimum, and it only works while the Fin AI agent is live. A team handling 4,000 conversations a month pays about $9,600 a year for Pro alone. UserInsight connects to Intercom, groups every conversation into product themes with their negative share, and costs from $119.50 a month for the whole account.
Unify · surface the why · traced to evidence
Teams that search for Intercom sentiment analysis usually want one of two things. Support leads want to spot the unhappy conversation before it turns into a cancellation. Product and CX leads want a monthly answer to a different question: which parts of the product generate the negative conversations, and did the last release make that better or worse.
Intercom now covers part of both jobs natively, with conditions most comparison pages skip. This page explains what CX Score and Topics Explorer actually measure, what they cost at real conversation volumes, read from Intercom's own help center and pricing in November 2026, and when a separate analysis layer is the cheaper and more useful buy. If your goal is only quality monitoring inside the inbox and Fin already runs your front line, the native route is a sound purchase and we say so below.
Traced to source evidence
No PII · GDPR-friendly
Why it works
What your team gets with Intercom sentiment analysis
Themes, not scores
Every Intercom conversation is grouped into a product theme, so negative sentiment points at a feature instead of a single chat.
No Fin required
Analysis runs on all your conversations whether or not Fin is live, and the price does not grow with every conversation.
Tied to behavior
Themes link to product usage, so you can rank problems by the accounts they affect and see if a release changed them.
What it handles
Unified, analyzed and surfaced, automatically
UserInsight unifies your sources, reads behavior and voice together, and surfaces the churn reasons, feature requests, friction steps and themes, each traced to the signals behind it.
- Connects to Intercom read-only, without changing inboxes, workflows or Fin
- Shows the negative share of every conversation theme
- Tracks whether a theme improves after a release
- Links each theme to the source conversations as evidence
- Joins Intercom themes to reviews, surveys and product usage
Top churn reason
Onboarding stalls before the first project
traced to 214 tickets + a 9% drop-off at onboarding step 3
Why UserInsight
One platform that fuses behavior and voice
Not an analytics tool that only shows the what, not a feedback repository that is blind to behavior. UserInsight joins both and surfaces the why, on aggregate consented data with no PII.
Unifies every source
Usage analytics, tickets, reviews, surveys and in-app feedback come together in one model, so behavior and voice finally live in the same place.
Surfaces the why
You do not write a query and wait. UserInsight tells you why churn moved and what to build next, ranked and quantified, the moment it changes.
Traced to evidence
Every insight links back to the specific tickets, reviews and events behind it, so you can click through and trust what you act on. No black box.
At a glance
Ways to get sentiment analysis on Intercom conversations, priced at 4,000 conversations a month
| Option | What you get | Price basis | About a year at 4,000 conversations a month | Best for |
|---|---|---|---|---|
| Intercom seats only, no add-on | Standard reports on volume, response time and CSAT, no CX Score and no Topics Explorer | Per seat, from $29 (Essential) to $132 (Expert) a month on annual billing | Seats only | Teams that only need operational reporting |
| Intercom Pro add-on with Fin live | CX Score from 1 to 5 with reasons, Topics Explorer, weekly Trends, Monitors | $0.20 per conversation, 500 conversation minimum ($100 a month), plus Fin from $0.99 per outcome | About $9,600 for Pro, plus your Fin outcomes | Quality monitoring for teams already running Fin |
| Export to your own model | Whatever your engineers build on the Conversations API | Engineering time plus model costs | Varies, ongoing upkeep | Teams with spare data engineering capacity |
| Dedicated feedback analytics vendor | Themes and sentiment across Intercom and other channels | Custom annual quote | Quote only | Enterprise CX programs |
| UserInsight | Every conversation grouped into product themes with negative share, trend and linked usage | Per account, from $119.50 a month billed annually | From $1,434 | Product and CX teams deciding what to fix |
Pro add-on pricing and CX Score requirements are from the Intercom help center and intercom.com/pricing, read in November 2026. Seat prices are from our October 2026 read of Intercom pricing. Yearly totals are our arithmetic. UserInsight does not replace Intercom, it reads from it, so for most teams the honest comparison is the Intercom plan they already pay for plus one analysis layer.
Does Intercom have sentiment analysis?
Yes, through CX Score. Intercom uses machine learning to rate each eligible conversation from 1 (negative) to 5 (positive), and a rating of 4 or 5 counts as positive in reports. Every score comes with CX Score reasons: answer quality, customer effort, strong emotion, product or service feedback, and policy feedback, each marked positive, negative or neutral. Intercom states there is no hidden weighting, the overall rating follows from the reasons that are not neutral.
The score is read from the customer's messages and the replies from Fin and your teammates. Internal notes, tags and SLAs are not used. A conversation qualifies when it is closed, user initiated, and has at least two customer replies and two agent or bot replies, so short one-question chats are never scored.
Three conditions decide whether you get it at all. CX Score requires the Pro add-on. It is only available to customers using Fin. And Intercom says it is not meant for assessing individual teammates, so it is a view of the customer experience, not an agent scorecard.
How much does Intercom sentiment analysis cost?
The Pro add-on is billed at $0.20 per conversation with a 500 conversation monthly minimum, which Intercom's pricing page shows as $100 a month. Fin has to be live as well, from $0.99 per Fin outcome. Seats are separate: $29, $85 and $132 per seat a month on Essential, Advanced and Expert with annual billing.
Because Pro is metered on conversations, the cost of sentiment rises with support volume rather than headcount. At 1,500 conversations a month Pro costs about $3,600 a year. At 4,000 it is about $9,600. At 10,000 it is about $24,000. Add Fin outcomes on top: if Fin resolves 800 conversations a month, that is another $9,504 a year, although most teams that buy Pro are already paying for Fin for its own sake.
That math is easy to defend when the value is operational, such as catching a spike of strong emotion during an outage. It is harder to defend when the question is analytical, such as which feature drove last month's negative conversations, because that needs one analysis across all conversations rather than a score on each one.
Topics Explorer and where it stops
Topics Explorer, also part of Pro, groups conversations into AI generated topics and subtopics and shows each topic's CX Score, Fin resolution rate and handling times. It is the closest Intercom gets to a theme report, and for a support operation it is genuinely useful.
It has limits a product team should know before buying. Intercom's own article says AI topics do not analyze sentiment themselves, the sentiment you see is the CX Score of the conversations inside the topic. Topics Explorer stays available only as long as Fin is live, it is offered in the US, EU and AU hosting regions, and the initial model is built from the past 90 days. Topics are also support categories by design (billing, login, refunds), not product areas.
The bigger gap is everything outside Intercom. A theme such as exports time out on large files shows up in Intercom conversations, app store reviews, NPS comments and the cancellation survey at the same time, and nothing in a help desk joins those or tells you whether the users writing in are your most active accounts.
Intercom conversation analysis for product teams
UserInsight connects to Intercom directly with a read-only connection and pulls new conversations without changing your inbox, workflows or Fin setup. It groups conversations into themes written in plain language, counts the customers behind each theme, shows the share of each theme that is negative, and links every theme back to the source conversations so nobody has to trust a summary they cannot check. It does not need Fin and it does not bill per conversation.
The part a help desk cannot do is the join. Import reviews, survey answers and usage data by CSV and each theme is tied to what those users did in the product, so a negative theme can be ranked by how many active or at-risk accounts it touches rather than by how loud it was. Personal data is not exposed in the analysis.
Pricing is per account. Starter is $119.50 a month billed annually and covers up to 120 AI analyses a month, and Growth is $359.50 with 400. A team handling 10,000 conversations a month pays the same plan price as one handling 1,000.
Support ticket analysis across Zendesk and Intercom
Customer sentiment analysis across tickets, reviews and surveys
When the native Intercom feature is the better buy
Buy Pro and skip a separate tool if Fin already handles your front line and your goal is operational: watching CX Score by topic during an incident, spotting conversations with strong emotion, or tracking whether Fin answers are hurting the experience. That work belongs inside Intercom, next to the conversations.
Add an analysis layer when someone outside support owns the outcome, typically a product manager, a head of CX or a founder, and they need a monthly answer to what is making customers unhappy and what would fix it. Teams without Fin have no native route at all, and teams with high volume often find a per account price cheaper than a per conversation meter.
If part of your support runs on Zendesk, the same logic applies, and we priced it in the same way on our Zendesk sentiment analysis page. Teams that compare dedicated vendors can see what Chattermill charges in our breakdown of its plans.
Good questions
Questions about Intercom sentiment analysis
Explore more
More ways teams find the why with UserInsight
Voice of customer software
Unify every customer voice and tie it back to what they actually do.
Learn moreCustomer feedback software
Collect feedback everywhere, then learn what it means in minutes.
Learn moreCustomer sentiment analysis
Track how customers feel, and why it is changing, in real time.
Learn moreStop guessing. See why users churn and what to build next.
Unify your usage data, feedback, tickets, reviews and surveys, and UserInsight surfaces the why, automatically. Aggregate and consented, with no PII.
Unifies usage, feedback, tickets, reviews and surveys · traced to source · no PII