How to use an Intercom bot to pre-qualify returns then create a ticket (and a simpler way)

If you’re running an e-commerce store, you know returns are just part of the deal. But managing the constant flood of requests can feel like a full-time job in itself. It's repetitive work that eats up your support team's day and pulls them away from customers who have more complex problems.

Automation is the obvious fix. The dream is to have a system that weeds out ineligible requests, gathers all the necessary info upfront, and then hands a neat, ready-to-go ticket to your team. While you can definitely build a workflow for this using tools like Intercom, it’s not always as simple as it looks on paper.

In this post, we’ll walk through how to set up an Intercom bot for this exact task. But more importantly, we’ll get real about the common snags you might hit, like wrestling with integrations and surprise costs, and show you a much more straightforward way to get it done.

Understanding Intercom's automation tools

Intercom’s world of automation is built around two main players: its AI Agent, Fin, and its more traditional, rules-based Custom Bots. They’re both designed to chat with website visitors, answer common questions, qualify leads, and send conversations to the right people.

Fin is the big-brain AI of the duo, capable of handling pretty complex conversations and solving issues all on its own. It's powerful, but it comes with a resolution-based price tag, which we'll talk more about in a bit. On the other hand, Custom Bots are what you'd typically use for a specific, step-by-step process like pre-qualifying returns. You build them with a set of rules and branching paths to guide a user from point A to point B.

Both of these tools are designed to work best inside the Intercom ecosystem. This is great if you’re all-in on their platform, but it can create some headaches if your bot needs to talk to other apps or pull in data from outside sources.

How to build a returns bot in Intercom

So, what does it actually take to build this bot? The process involves sketching out the conversation, building the logic in Intercom, and then setting it live. Let's walk through it.

Step 1: Map out the conversation flow

First things first, you need to script out the entire conversation. Think about what information you absolutely have to collect before a return ticket even gets created.

Your flow will probably look something like this:

1. The bot kicks things off with a friendly greeting and confirms the customer wants to start a return.

2. It then asks for the basics, like an order number and the customer's email.

3. Next, it asks why the item is being returned (e.g., wrong size, damaged, just changed my mind). This is super useful data for tracking product issues later.

4. Here’s the tricky part: the bot should check the request against your actual return policy (like, was the order placed in the last 30 days?). This is often the hardest step to automate within a closed system.

5. If everything looks good, the bot creates a support ticket and lets the customer know what happens next.

6. If the return isn't eligible, the bot needs to explain why, point them to your policy, and maybe offer a way to talk to a human if things get complicated.

Step 2: Set up the custom bot

With your script in hand, you’d head into Intercom’s Custom Bot builder. It’s a visual editor where you create the different conversational paths based on the map you just made. You’ll set up rules that trigger specific questions based on how the customer answers.

A view of Intercom's visual workflow builder, where a user can set up a custom bot for returns.

The grand finale of this workflow is the action: creating a new ticket. You can tell the bot to automatically assign this ticket to your returns team, add tags like "Return Request," and set its priority. The goal is that when an agent finally sees the ticket, they have all the context they need without a single back-and-forth email.

Step 3: Train and deploy the bot

Once it's built, you have to test it. Click through all the paths to make sure there are no dead ends or weird loops. When you feel good about it, you can push it live. You'll want to keep an eye on how it’s doing, though it can be tough to know how it'll handle every real-world scenario without a proper testing environment.

Challenges with the native Intercom setup

Building a a basic returns bot in Intercom is one thing. Making it smart, efficient, and not a massive headache is another. Teams often run into a few common hurdles that can derail the whole project.

Unpredictable costs

To get the most intelligent, human-like responses, you'll want to lean on Intercom’s Fin AI. The catch? Intercom charges $0.99 for every resolution Fin handles. While that might not sound like much, think about your peak seasons. After a holiday sale, you could be processing hundreds or even thousands of returns. Those 99-cent charges add up incredibly fast, leaving you with a bill that’s hard to predict and punishes you for growing.

Limited access to external knowledge

This is probably the biggest roadblock. A standard Intercom bot is basically stuck in its own little world. To трули pre-qualify a return, your bot needs to see the order date from Shopify, check the eligibility rules in your Google Doc, or look up a customer’s purchase history in your CRM.

Getting an Intercom bot to access that kind of real-time, external info usually means custom API work, which means roping in developers. Without that, your bot isn't really "qualifying" anything; it's just asking a list of questions.

This diagram illustrates how a native bot like Intercom's has limited knowledge sources compared to an agnostic AI.

The "rip and replace" mindset

Intercom’s automation is at its best when your entire support world is Intercom: their help desk, their knowledge base, their messenger. If your team is already happy using a different help desk like Zendesk or Freshdesk, you’re stuck. You either have to migrate your entire operation or live with a clunky, disconnected system. It doesn’t just layer on top of the tools you already know and love.

Difficult to test safely

What happens if your bot is configured wrong? You could end up annoying customers with broken logic or creating a tidal wave of messy tickets for your team. The risk of deploying a bot without deep testing is you won’t know how it actually performs until it’s live with real customers. Intercom doesn’t have a built-in simulation mode that lets you test your bot against thousands of your past tickets, so you’re more or less flying blind.

Intercom's pricing model explained

To really get the cost issue, it helps to see how Intercom’s plans are put together. They charge per person, but the powerful AI automation comes with that extra usage-based fee on top.

To use Fin, their AI Agent, you need at least one paid seat on a plan. Then, on top of that monthly seat cost, you pay for each successful resolution.

Plan	Price per Seat/mo (Annual)	Fin AI Agent Cost	Key Features
Essential	$29	$0.99 per resolution	Shared Inbox, Help Center, Basic Reports
Advanced	$85	$0.99 per resolution	Workflows, Multiple Inboxes, Round Robin
Expert	$132	$0.99 per resolution	SSO, SLAs, Multibrand Support

That $0.99 fee might look small on a pricing page, but for a high-volume task like returns, it creates a variable cost that’s a nightmare to budget for.

A simpler way to automate returns with eesel AI

The headaches that come with native tools have opened the door for a smarter, more flexible approach. Instead of a platform that asks you to "rip and replace" everything, eesel AI works as an intelligent layer that plugs right into the tools you already have, including Intercom, Zendesk, and Shopify. It’s built to solve the exact problems we just covered.

Go live in minutes with one-click integrations

Getting started with eesel AI is refreshingly simple. You can connect it to your help desk and knowledge sources in a few clicks, no complicated API projects needed. It’s a truly self-serve platform, which means you can set up, configure, and launch your returns bot all on your own, without having to sit through a mandatory sales demo just to get access.

Unify your knowledge for smarter qualification

This is where eesel AI really changes the game. You can train it on all your scattered knowledge at once: your official return policy living in a Google Doc, your product details in Shopify, and even the outcomes from thousands of past return tickets.

This means your bot doesn't just follow a rigid script. It actually understands the nuances of your policies and can accurately pre-qualify returns based on real-time information. It can answer questions like, "Can I return a final sale item if it arrived damaged?" with an answer that’s actually correct for your business.

Test with confidence using simulation mode

Remember that fear of deploying a bot without knowing how it will behave? eesel AI solves this with a powerful simulation mode. Before your bot ever chats with a real customer, you can run it against thousands of your historical support tickets.

You’ll see exactly how it would have answered, what its resolution rate would have been, and where you might have gaps in your knowledge. This lets you tweak its performance and roll it out with total confidence, knowing it’s ready for whatever your customers throw at it.

Take control with customizable actions and predictable pricing

With eesel AI, a bot can do more than just chat. You can set up custom actions that let it look up an order status in Shopify, check a customer's loyalty tier in your CRM, or create a ticket with specific custom fields filled out.

And best of all, the pricing is completely transparent and predictable. eesel AI’s plans are based on your overall usage, with no per-resolution fees. Your bill won't suddenly shoot through the roof after a busy season, so you can scale your automation without any financial surprises.

Automate smarter, not harder

Automating your returns process is one of the biggest wins you can get for your support team's sanity and efficiency. While tools like Intercom offer a place to start, they can also introduce new headaches, from unpredictable costs to frustrating technical limits.

Modern AI platforms like eesel AI offer a more flexible, powerful, and cost-effective way forward. By working with the tools you already have and giving you the power to test everything thoroughly, they let you build smart workflows that actually solve problems instead of creating new ones. You can free up your team, give customers faster answers, and get back to focusing on growing your business.

Ready to see how easily you can automate returns and other repetitive support tasks? Try eesel AI for free or book a demo to see the simulation mode in action.