A complete overview of the Intercom chatbot in 2025

Intercom is a huge name in customer communication, and for good reason. Their AI chatbot, Fin, promises to handle support tickets, answer questions 24/7, and free up your team to do more important work. On paper, it sounds pretty great.

But as a lot of teams have found out, the reality can be a bit messy. It’s not wild to see your bill double overnight after turning on the AI, all because of a pricing model that charges you for every single problem your chatbot solves. This begs the question: should your support tool penalize you for being good at your job?

We don’t think so. This guide gives you the full picture of the Intercom chatbot, what it does well, where it stumbles, and how you can get all the good parts of AI without the unpredictable costs.

What is the Intercom chatbot?

The Intercom chatbot, which they call the Fin AI Agent, is Intercom’s built-in AI tool for handling frontline customer support. Its main purpose is to understand what customers are asking and give them automated answers, hopefully resolving their issues without a human agent ever getting involved.

A screenshot of the Intercom chatbot interface, showing how it interacts with users.

You can think of it as a digital team member that hangs out in your Intercom messenger and helpdesk. It’s built on some powerful tech, including OpenAI’s GPT-4, and you can train it on your company’s own knowledge, like your help center articles and website content.

The goal is simple enough: cut down customer wait times, offer support around the clock, and let your human agents focus on the really tricky conversations that need a personal touch. When it’s working right, it can deflect a ton of common questions, from "How do I reset my password?" to "What’s your return policy?". It’s a big piece of Intercom’s move toward being an "AI-first" platform.

Key features of the Intercom chatbot

Before we get into the tricky parts, let’s cover what the Intercom chatbot is actually designed to do. It’s a solid tool with a few key features that make it a go-to for teams already using the Intercom ecosystem.

Training on your knowledge base

At its heart, Fin learns from the information you feed it. You can connect it to your public help center, upload files like PDFs, or point it to certain pages on your website. The chatbot then uses all that content to piece together answers for customers. For instance, if someone asks about shipping times, Fin will look for the right article in your knowledge base and give them a summary.

An image displaying the interface for connecting various knowledge sources to the Intercom chatbot.

But, as many users have pointed out, the quality of its answers is directly tied to how clean and complete your knowledge base is. If your docs are outdated or missing information, the chatbot’s performance is going to take a hit.

Workflow automations and custom flows

Intercom has a visual flow builder that lets you create custom, step-by-step chats. These aren’t just for answering questions; you can build workflows to qualify new leads by asking them a few questions, grab customer info before a handoff to a live agent, or point users to the right resource.

A screenshot of the Intercom visual flow builder, which allows for the creation of custom chatbot workflows.

These custom flows give you more control over the conversation, letting you create guided paths for common problems instead of just hoping the AI figures it all out on its own.

Proactive support

The Intercom chatbot isn’t just reactive. You can also set it up to proactively message visitors on your website. You can create triggers for the chatbot to pop up based on what a user is doing, like how long they’ve been on a page or a specific button they’ve clicked. This is useful for things like announcing a sale, offering an in-app tour for new users, or sending messages to bring customers back. It turns the chatbot from just a support tool into something you can use for marketing and onboarding, too.

The Intercom chatbot proactively engaging with a website visitor to offer assistance.

The hidden costs: Pricing

Alright, here’s where things get complicated. The biggest complaint about the Intercom chatbot usually isn’t about its features, but its price tag. It can be confusing, hard to predict, and, frankly, very expensive.

Intercom’s pricing has two main components:

1. A per-seat fee: You pay a monthly subscription for each human agent on your team. This goes from the Essential plan at $39/seat/month to the Advanced plan at $99/seat/month.

2. A per-resolution fee: On top of the seat license, you also pay $0.99 every time Fin successfully resolves a conversation.

That second part is the kicker. What exactly is a "resolution"? Intercom says it’s any conversation where the customer confirms their problem is solved or just stops responding after the AI’s last message. That gray area means you’re often paying for resolutions that might not have actually been helpful.

Let’s look at a real-world example. Say your chatbot handles 5,000 customer chats in a busy month. That’s an extra $4,950 on your bill, on top of what you’re already paying for your agents. This is exactly how a $4,000 monthly bill can balloon to $9,000, a situation

.

The main issue is that this model punishes you for growing. The more customers you get and the more they reach out for help, the higher your bill climbs. You end up in a weird spot where you might be hesitant to let your support tool do its job because you’re worried about the cost.

This pay-per-resolution model isn’t the only way. Modern AI platforms like eesel AI offer predictable, flat-rate pricing. You know exactly what you’ll pay each month, which lets you scale your support without worrying about your costs spiraling out of control.

Feature	Intercom Chatbot (Fin)	eesel AI
Pricing Model	Per-seat + Per-resolution ($0.99/resolution)	Flat monthly/annual fee
Cost Predictability	Low (scales with every ticket)	High (fixed price)
Hidden Fees	High potential for surprise resolution charges	None, just transparent plans
Best For	Teams with low and predictable support volume	Teams who want to scale support without scaling costs

Common headaches with the Intercom chatbot

Pricing aside, teams often bump into a few other practical problems when they start using Intercom’s native AI.

Your help center has to be perfect

We mentioned this earlier, but it’s a big one: Fin is only as smart as the knowledge you give it. If your help center isn’t perfectly written, organized with military precision, and constantly updated, the chatbot’s performance will be, as one user described it, "below average." This creates a huge upfront project for teams just to get their documentation ready, not to mention the ongoing work to keep it all accurate.

A view of the Intercom knowledge base, where articles and documents are managed to train the Intercom chatbot.

Limited control over your rollout

Switching on an AI chatbot can feel like a leap of faith. With Intercom, it’s tough to test how the chatbot will actually perform (and what it will cost you) without just flipping the switch and hoping for the best. You can’t easily run a simulation on your past conversations to see how it would have done. This "all-or-nothing" approach feels risky, especially when every single resolution adds to your monthly bill.

Why knowledge is siloed within Intercom

Intercom’s chatbot mostly learns from the help center content you feed it. But what about the most valuable source of knowledge your company has? The thousands of past support conversations where your expert agents have already solved real customer problems. That entire history of solutions, plus all the internal knowledge you have stored in places like Confluence, Google Docs, or Notion, is completely ignored.

A smarter Intercom chatbot: Add a dedicated AI layer to Intercom

Instead of being locked into an expensive and rigid system, what if you could just add a smarter, more flexible AI layer on top of the Intercom setup you already have? That’s the idea behind eesel AI. It’s designed to work with your helpdesk, not trap you inside a walled garden.

Access all your knowledge sources

eesel AI connects to all of your company’s knowledge, not just one polished help center. It can train on your past Intercom tickets, learning directly from how your best agents have solved problems before. It also integrates with Confluence, Google Docs, Notion, Shopify, and dozens of other tools, making sure the AI has access to the same information your team does. This helps it give answers that are more accurate and sound more like your brand right from the start.

A workflow diagram illustrating how a dedicated AI layer can pull information from multiple sources compared to a native chatbot.

Go live in minutes with full control

Forget about long sales calls and mandatory demos. eesel AI has a completely self-serve setup you can get running in minutes. The one-click integration with Intercom is about as smooth as it gets.

Best of all, you can take the risk out of the whole process with a powerful simulation mode. Before the AI ever talks to a real customer, you can run it on thousands of your past Intercom tickets. You’ll get a clear forecast of its resolution rate and see exactly how it would respond to real questions. This gives you the confidence to roll it out slowly, maybe starting with just a few common ticket types and expanding from there.

Get predictable pricing you can count on

This is the big one. With eesel AI, you pay a flat, predictable monthly fee. There are no per-resolution charges, ever. You can automate 100 tickets or 10,000 tickets, your bill doesn’t change. This finally lines up the cost of your AI tool with your business goals, letting you scale up your support without fearing the end-of-month invoice.

What’s the verdict on the Intercom chatbot?

The Intercom chatbot is a decent native tool that can work for teams with a flawless help center and a low, steady stream of support tickets. For them, the convenience of having it all in one place might be worth it.

However, for most growing businesses, its reliance on perfect documentation and its wild, per-resolution pricing model create some serious hurdles. For teams that need more flexibility, want to use all their company knowledge, and need predictable costs, adding a dedicated AI layer like eesel AI is a much more powerful and sustainable way to automate support.

This video shows how you can build a custom AI chatbot for Intercom to get more control and better results.

Ready to try a better Intercom chatbot?

See how AI can automate your support without the surprise costs. Sign up for a free eesel AI trial and run a simulation on your own tickets in just a few minutes.