How to use Intercom Fin to collect missing order info then hand off to agent

If you work in customer support, you know the drill. A message pops up: "Where's my order?" And... that's it. No order number, no email, nothing. So begins the back-and-forth that eats up time for both your agent and the customer. This is exactly the kind of thing tools like Intercom’s Fin AI agent promise to fix by handling that initial info-gathering for you.

In this guide, we'll get into the nitty-gritty of setting up Intercom Fin to collect missing order info then hand off to an agent. We'll go through the official Intercom way, but I’ll also be real about some of the tricky parts and the costs. Then, we’ll look at a more straightforward way to get this type of automation up and running.

What you'll need to get started

Alright, before you roll up your sleeves and dive into Intercom, let's make sure you have everything you need. Getting this workflow running isn't as simple as flipping a switch; you'll need to connect a few different tools and have the right permissions.

• An Intercom Plan with Fin AI Agent: First off, you need a subscription that includes Fin. It’s usually on their Essential, Advanced, or Expert plans, but keep in mind that resolutions are billed on top at $0.99 a pop.

• Admin Access: You'll need to be an admin in your Intercom workspace to get into the settings for Fin, Data Connectors, and Tasks.

• API Access to Your E-commerce Platform: Fin needs a way to peek into your store's backend to check on orders. This means you’ll need an API key for your e-commerce platform, whether that’s Shopify, BigCommerce, or your own custom system.

• Technical Familiarity (or a Developer): Let's be honest, setting up the connection to your store involves some technical bits. While Intercom tries to make it accessible, you’ll be dealing with terms like API endpoints, headers, and JSON. If that sounds like another language, you might need to grab a developer for help.

How to set up the Intercom Fin workflow

Okay, let's get into it. The whole setup in Intercom boils down to two main components: Data Connectors, which let Fin talk to your store, and Fin Tasks, which tell Fin what to say and do.

Step 1: Connect your store with a data connector

First things first, you need to give Fin a way to look up orders. Think of a Data Connector as a special hotline between Intercom and your e-commerce store. You’re essentially telling Fin, "When you need order info, use this number to call the store's database."

1. Head to Fin Tasks & Data Connectors: You'll find this in your Intercom workspace settings.

2. Create a New Data Connector: Give it a clear name that you'll remember, like "Fetch Order Details."

3. Configure the API Call: This is the most technical part. You have to paste in the API endpoint from your store that's used for looking up orders. You'll also need to handle the security side of things, which usually means adding your store's API key to the request header so it knows the request is legit.

4. Define Inputs and Outputs: Here, you'll map out the data. You need to tell Intercom what info Fin needs to send over (like an "order_id") and what info you want to get back (like "order_status").

This step is make-or-break, and it's honestly where a lot of people get bogged down. If your store's API has some quirks or the data needs to be cleaned up before Fin can use it, things can get complicated fast.

A screenshot showing the Intercom Fin interface, relevant to the process of setting up Intercom Fin to collect missing order info then hand off to an agent.

Step 2: Create a Fin task

Now that Fin can look up orders, you need to teach it how to have the conversation. That's what a Fin Task is for. You'll write out instructions in plain English to guide Fin through the chat.

1. Start a New Fin Task: Call it something descriptive, like "Find and Check Order Info."

2. Set the Trigger: Tell Fin when to jump in. You can use common customer questions like "where is my order?" or "can you check my order status?" as triggers.

3. Write the Instructions: This is where you script out the conversation logic. You'll write out commands for Fin, including what to do when things don't go as planned. It might look something like this:

   • "Ask the customer for their order number."

   • "If they give you one, use the 'Fetch Order Details' Data Connector to look it up."

   • "If you find the order, tell the customer what you found and ask them to confirm."

   • "If you can't find the order, let them know and ask them to check the number again."

   • "If the customer says they don't have the order number, ask for their email address instead."

Getting this part right usually involves a bit of trial and error. You have to think like a customer and cover all the weird ways they might respond so the AI doesn't get stuck.

Step 3: Set up the agent handoff

This is the final and most important part: making sure the conversation gets to your team smoothly. The idea isn't for Fin to solve every single problem, but to tee up the conversation so your agent can swoop in and solve it right away.

Inside your Fin Task instructions, you'll add a final command to pass the conversation over to a human. You can trigger this once the order info has been successfully collected.

Example Instruction:

"After the customer confirms their order details are correct, Escalate to: Support Team."

You can also add fallback rules, like handing off the chat if the customer asks to "talk to a human" or if Fin fails to find the order after a couple of tries. This makes sure no one gets trapped in an automation loop. Once it's all set up, Fin will handle the tedious part, and the ticket will land in your team's queue with all the key details attached.

Common challenges with this approach

So, the process we just walked through does work. But it's worth knowing about a few common bumps in the road:

• Setup Complexity: Let's not sugarcoat it: this can be complicated. You need to be comfortable with API calls, and writing the natural language instructions for Fin Tasks often takes a lot of fine-tuning to get them to work reliably.

• A Lack of Confident Testing: One of the more stressful parts is that it’s hard to know if your setup really works until it's live with actual customers. The "test in production" reality can lead to some awkward conversations if your logic has a hole in it.

• Unpredictable, Per-Resolution Pricing: This is a major sticking point for many teams. Intercom's $0.99 per resolution fee sounds simple, but a "resolution" can just mean any conversation where the customer didn't explicitly ask for a human. Your bill can swing wildly from one month to the next, which makes budgeting a headache. Ironically, the better your automation performs, the higher your bill gets.

A simpler, more powerful way with eesel AI

If those challenges sound a bit too familiar, and you're looking for automation without the technical headaches and surprise bills, it's worth looking at a different approach. A tool like eesel AI was built to solve these exact problems.

Here’s how you could build the same "where is my order?" workflow with it:

Go live in minutes with one-click integrations

Instead of wrestling with APIs, eesel AI has a gallery of over 100 integrations. You can connect to Intercom, Shopify, Google Docs, Confluence, and other tools in a matter of seconds. All the messy authentication and data mapping stuff is handled for you.

A screenshot showing eesel AI's one-click integrations, a simple alternative to the complex setup of Intercom Fin to collect missing order info then hand off to an agent.

Test with confidence using a powerful simulation mode

This is a huge relief for most teams. Before your AI agent talks to a single customer, you can run a simulation on thousands of your past tickets. eesel AI will show you exactly how it would have replied and what actions it would have taken, giving you a solid forecast of its performance. This lets you tweak your logic and go live feeling good about it.

A view of eesel AI's simulation mode, which allows for testing before deploying a workflow like Intercom Fin to collect missing order info then hand off to an agent.

Total control and predictable pricing

With eesel AI, you use a visual workflow builder to define exactly how your AI behaves. You can set up precise rules for which tickets get automated and which get passed to your team. And maybe best of all, the pricing is straightforward. You pay a flat monthly fee, not per resolution. That means your costs stay the same whether your AI handles 100 tickets or 10,000, so you can scale up your success without blowing your budget.

Making the Intercom Fin workflow work for you

Automating routine questions like "where's my order?" is a huge win for any support team. It frees up your agents to focus on more complex issues and gives customers faster answers. While you can definitely use Intercom Fin to collect missing order info then hand off to an agent, the path can be technically demanding, hard to test safely, and comes with a pricing model that punishes you for being successful.

For teams that just want a solution that’s quick to set up, safe to roll out, and easy to budget for, a platform like eesel AI is a clear alternative. By focusing on a self-serve setup, proper simulation testing, and predictable pricing, you can build the automations you need in minutes and scale your support without stressing about the bill.

Ready to see how easy it can be? You can start a free trial of eesel AI and see if you can get your first automation running in the next five minutes.