So, you’re looking into using Intercom’s Fin AI to help out with customer support. It’s a tool that gets a lot of attention, promising to answer customer questions in a flash and free up your team for the trickier stuff. But let’s be honest, going from signing up to actually having a helpful AI assistant isn’t always a straight shot. Where do you even begin? How do you teach it, tell it how to act, and launch it without causing a headache for your customers?
This guide will walk you through exactly how to use Fin AI, one step at a time. We’ll cover everything from feeding it the right information to building out custom workflows and, just as importantly, avoiding the common traps people fall into. By the end, you’ll have a clear plan for getting Fin up and running.
What you’ll need to use Fin AI
Before we jump into the setup, let’s get our ducks in a row. Fin AI isn’t a standalone tool you can just plug in anywhere; it lives and breathes inside the Intercom world. That tight connection is part of its power, but it also means you’re committing to their platform.
Here’s a quick checklist of what you need to have in place:
-
An active Intercom subscription: Fin AI is an add-on, so you’ll need a main Intercom plan to even get access to it.
-
Admin access: You’ll need to be an administrator in your Intercom workspace to get to the settings where all the magic happens.
-
A knowledge base: An AI is only as smart as the information you give it. Fin learns from your support content, so having a well-organized and up-to-date help center is a must. If it’s not in good shape, you might want to tidy it up first.
-
A clear idea of what you want to automate: Are you trying to deflect simple "how-to" questions or handle more complex account queries? Knowing what you’re aiming for from the start will make the whole process a lot less painful.
How to use Fin AI in 6 steps
Alright, let’s get into the nitty-gritty. We’ve broken the process down into six main stages, from teaching Fin the basics to letting it talk to your customers for the first time.
Step 1: Train Fin AI on your support content
First things first, Fin needs to learn. Think of it like a new hire on their first day; it doesn’t know anything until you show it the ropes. Fin’s main source of knowledge is your Intercom Articles (your help center). You’ll need to head into the Fin settings and point it to the content it should use.
Make sure your articles are clear, accurate, and actually answer the questions your customers are asking. If your knowledge is spread out all over the place, you’ll likely need to pull it all into Intercom to get started. This part is really important because Fin’s ability to give good answers is directly tied to the quality of the info you feed it.
A screenshot showing how to connect knowledge sources in Intercom, which is the first step in how to use Fin AI.Step 2: Configure Fin AI’s personality and guidance
Next up, you need to decide how you want Fin to sound. You don’t want a stiff, robotic assistant greeting your customers. In the settings, you can customize Fin’s tone of voice and give it "guidance" on how to handle certain topics.
For instance, you can tell Fin things like:
-
Always keep the tone friendly and understanding.
-
Never make promises about when a new feature will be released.
-
If a customer mentions the word "refund," immediately pass the conversation to a human.
This is your chance to make the AI feel like a genuine part of your team and brand. Getting the voice right goes a long way in building trust with your customers.
This image shows the chatbot guidance settings in Intercom, an important part of learning how to use Fin AI.Step 3: Set up basic deployment and handover rules
Intercom has a "Simple deploy" option to get you going without too much fuss. Here, you can set the ground rules for when Fin should jump in and who it should talk to (like new visitors only, or just logged-in users).
The most critical piece here is figuring out what happens when Fin gets stumped. A smooth handoff to a human agent is non-negotiable. You can set it up to automatically send the chat to the right team (like Sales or Support) and let the customer know when they can expect a reply. An AI that leads to a dead end is often more frustrating than no AI at all, so make sure this escape hatch works perfectly.
Step 4: Build advanced workflows for complex queries
When you’re ready to move beyond basic Q&A, you’ll need to dive into Intercom’s Workflows. This is a visual tool where you can map out conversation paths with custom logic. You can drop a "Let Fin answer" block into a workflow to have the AI take over at a specific moment.
For example, you could build a workflow that first asks a customer what they need help with ("Billing" or "Technical Issue") and then lets Fin take a crack at answering before sending it to the right person. These workflows are powerful, but be warned: they can get complicated fast and require a good amount of planning to make sure you’ve covered all your bases.
An example of a complex workflow built in Intercom's visual editor, demonstrating an advanced step in how to use Fin AI.Step 5: Automate actions with Fin tasks
Beyond just answering questions, you can teach Fin to do things using a feature called "Tasks." This lets Fin handle processes that have a few steps, like canceling an order or updating a shipping address, by connecting to your other systems.
Setting up Tasks is definitely a more advanced step. It’s a mix of clicking buttons and understanding how your APIs work. This is where you can get some serious automation done, but it often means you’ll need to pull in a developer or someone from your technical team to help get it hooked up correctly.
Step 6: Test and roll out your Fin AI agent
Finally, before you set Fin loose on all your customers, you have to test it. A lot. Intercom gives you a preview mode where you can chat with Fin yourself and see how it responds to different questions.
It’s also a good idea to roll it out slowly. Maybe start by turning it on for just 10% of your visitors or only on your pricing page. Keep a close eye on the analytics. What’s its resolution rate? Are customers happy with the answers? What questions is it failing on? Use that feedback to go back and improve its training content and guidance.
The preview mode in Intercom for testing Fin AI's responses, a crucial final step in how to use Fin AI.Common mistakes to avoid
Putting an AI agent in place isn’t a "set it and forget it" task. While the steps above will get you started, here are a few common tripwires to look out for.
Underestimating the setup time
While Intercom’s "Simple deploy" can get you started quickly, unlocking Fin’s real power with custom Workflows and Tasks takes time and careful planning. Many teams think it’ll be a quick weekend project, only to find it turning into a multi-month endeavor.
It’s a common story with tools that are deeply integrated into one platform. In contrast, tools like eesel AI are designed so you can get up and running on your own, often going live in minutes. You can connect it to your helpdesk with a single click and start automating tickets without getting lost in a workflow builder.
Forgetting about the total cost
Fin AI has a pay-per-resolution price tag ($0.99 for every issue it solves, at the time of writing). That might sound simple, but it can make your costs unpredictable, especially when you have a busy month. As your support volume goes up, so does your bill, which can make budgeting a real challenge.
A flat-rate plan can be much more predictable. For instance, eesel AI has transparent pricing with monthly or annual plans that don’t charge you for every single resolution. That way, you’re not getting penalized for having a successful month of automation.
Getting locked into one system
Fin AI only works with Intercom. Period. If your team ever decides to switch to another helpdesk like Zendesk, Freshdesk, or Jira, all the time and effort you invested in Fin is gone. It’s a bit like buying a coffee maker that only takes one specific, expensive brand of pods.
A more flexible approach is to use an AI tool that works with the tools you already use. eesel AI plugs into all the major helpdesks. This lets you get great automation today without having to move your whole support operation, and it keeps you free to switch tools down the road if you need to.
Get powerful AI automation without the complexity
Setting up Fin AI can definitely level up your customer support, but as we’ve seen, it takes real configuration work, continuous upkeep, and a full commitment to the Intercom platform.
But what if you could get the same (or even better) results with a lot less effort?
eesel AI is built for teams that want smart, customizable AI without the long setup times and vendor lock-in. You can get started for free and be up and running in minutes, not months.
Here’s how eesel AI makes automation simpler:
-
It works with your existing tools: eesel AI integrates smoothly with Zendesk, Freshdesk, Intercom, Jira, Slack, and more. No need to change your helpdesk.
-
It’s truly self-serve: You can set up, configure, and launch it all on your own. No need to sit through mandatory sales demos just to get started.
-
You can test with confidence: Use the simulation mode to test the AI on thousands of your real past tickets. You’ll know exactly how it will perform before it ever talks to a live customer.
-
The pricing is transparent: Our flat-rate monthly plans mean your bill won’t suddenly spike just because the AI had a great month.
If you’re looking for a faster, more flexible, and more predictable way to automate your customer support, give eesel AI a look.
Start your free eesel AI trial today and see for yourself how quickly you can get your frontline support automated.
Frequently asked questions
To start, you’ll need an active Intercom subscription, admin access, and a well-organized knowledge base. Then, you’ll train Fin AI on your support content and configure its personality and tone within the Intercom settings.
Your knowledge base is critically important as Fin AI learns directly from your Intercom Articles and public web pages. The quality and accuracy of your content directly determine Fin’s ability to provide helpful and reliable answers to customers.
When Fin encounters a question it can’t answer, a smooth handoff to a human agent is essential. You can configure automatic routing to the appropriate team and set expectations for the customer’s wait time within Intercom’s deployment settings.
Advanced workflows in Intercom allow you to map out custom conversation paths with specific logic, integrating Fin to answer specific queries at certain points. Tasks enable Fin to perform multi-step actions, like updating an order, by connecting to other systems, though this often requires technical assistance.
You should utilize Intercom’s preview mode to extensively chat with Fin and assess its responses. Rolling it out slowly, perhaps to a small percentage of visitors, and closely monitoring analytics will help you refine its training and guidance based on real-world performance.
Common pitfalls include underestimating the setup time required for advanced features and facing unpredictable costs due to Fin AI’s pay-per-resolution pricing model. Additionally, there’s the risk of vendor lock-in since Fin AI operates exclusively within the Intercom platform.
Fin AI is deeply integrated into, and exclusively works with, the Intercom platform. This means that any investment in Fin AI would be lost if your team decides to switch to a different helpdesk, limiting its flexibility in a broader support ecosystem.
