A practical guide on how to use Intercom workflows to tag conversations for CSAT analysis

Getting a bunch of CSAT scores back is a good feeling. You see the smiley faces, you see the frowny faces, and you get a general sense of how things are going. But let's be honest, a simple score only tells you what customers feel, not why.

To get to the good stuff, the insights you can actually act on, you have to dive into their comments and start categorizing the feedback. Trying to do this by hand is a massive time sink, and worse, it often leads to messy, inconsistent data because everyone tags things a little differently.

The good news? You can automate a good chunk of this right inside Intercom. This guide will walk you through exactly how to set up Intercom workflows to tag conversations for CSAT analysis. We'll cover building the workflow from the ground up, and then we'll look at how you can use AI to pull out much deeper insights, saving your team from hours of manual digging.

What you'll need to set up this workflow

Before we jump in, let's get our ducks in a row. Setting this up is pretty straightforward, but you’ll want to have a few things ready to go.

• An Intercom account: This might seem obvious, but you’ll need a plan that includes the Workflows feature. A quick look at Intercom's pricing page shows this is available on their Advanced plan and up.

• Admin access: You’ll need the right permissions to create and manage workflows in your Intercom workspace. If you don't have them, you'll need to find the person who does.

• A basic tagging strategy: It helps to think about what you want to learn ahead of time. You don't need a perfect system, just a starting point. Simple tags like "csat-positive" and "csat-negative" are great. You could also add a few for common topics you already know about, like "pricing-issue" or "product-feedback".

Step-by-step: Building the Intercom workflow

Alright, let's build this thing. The goal is to create a system that automatically sends a CSAT survey when a conversation closes and then tags that conversation based on the customer's rating.

Step 1: Create a new workflow from scratch

First things first, head over to the Workflows section in your Intercom dashboard. Find and click the + New Workflow button. It'll ask you what you want to do, and you're going to choose to build your workflow from scratch. This gives you total control over how it works, which is exactly what we need.

Step 2: Set up the trigger

Every workflow needs a "go" signal. For this one, we want it to kick off right after a support chat ends.

1. For the trigger Category, pick During conversation only.

2. Then, choose the specific trigger: If teammate changes the conversation state.

3. In the settings for that trigger, select the Closed state. This little detail is important, it makes sure the workflow only runs when a person on your team closes a conversation, not a bot.

Step 3: Add the CSAT rating request

Once the trigger is set, the very next thing we want to do is ask for feedback. It’s best to do this immediately while the conversation is still fresh in the customer's mind.

1. Click the + Add Step button right after your trigger.

2. Select the action Ask for conversation rating. This is Intercom's built-in CSAT tool, so it plays nicely with everything else.

Step 4: Create branches

This is where the automation gets smart. The workflow needs to pause, wait for the customer's rating, and then take a different path depending on whether the feedback was good or bad. This is what makes the whole thing useful.

1. Click on the "Rate Your Conversation" step you just added.

2. You'll see a toggle for ‘Wait for customers to give a rating before continuing the Workflow’. Flip that on.

3. After that wait step, click + Add Step and choose Branch. This lets you create different paths based on the rating.

Step 5: Configure branches and apply tags

Now it's time to set the rules for each path and tell Intercom which tag to use. We'll keep it simple with a positive and a negative branch.

• Branch 1: Negative Feedback

  • Set the condition to "If Conversation Rating is less than Ok".

  • Inside this branch, click + Add Step and pick Tag conversation.

  • Choose a tag you already created or make a new one, something like "csat-negative-feedback". You could even add another step here to ping a manager in a specific Slack channel.

• Branch 2: Positive Feedback

  • This will be your "Else" path. The condition should be "If Conversation Rating is More than or equal to Ok".

  • Add the Tag conversation action and use a tag like "csat-positive-feedback".

You can always make this more complex later by adding a third branch for neutral ratings if you want to get more granular.

Step 6: Set your workflow live

Once your branches and tags are all set up, give the whole thing a quick once-over. Does the logic make sense? Does it flow from the trigger to the final tag? If it all looks good, hit the Set it live button. And that's it! Your workflow is now active and will start tagging conversations as the ratings roll in.

A complete view of the Intercom workflow builder, showing the trigger, actions, and branches for CSAT analysis.

Beyond the basics of tagging

Automating basic tagging is a fantastic first step. Seriously, it’s way better than doing nothing. But if you stop here, you'll eventually run into a few roadblocks that keep you from seeing the full picture of customer satisfaction.

The trouble with group chats

One of the first quirks teams notice is that Intercom doesn't send CSAT surveys for conversations with more than one participant. They do this on purpose to avoid confusion over who should be rating the chat, but it means you're missing out on feedback from any ticket where you had to loop in a colleague. As people have pointed out in the Intercom Community, your workflow will fire, but the survey never actually goes out, leaving a frustrating blind spot in your data.

Why basic workflows don't tell you "why"

A "csat-negative-feedback" tag is a clear signal that a customer was unhappy. But was it because of a product bug? A missing feature? A pricing issue? Or maybe just a slow response? To find the "why," your team still has to open up every single one of those tagged conversations and read through them manually. This is the exact soul-crushing work you were trying to automate away. It's slow, subjective, and simply doesn't work once your ticket volume starts to climb.

Why complex workflows get messy, fast

So, you think, "I'll just add more branches for keywords!" You could try building out a workflow with dozens of branches looking for specific words in the conversation. But this approach quickly turns into a tangled web of rules that is a nightmare to build, test, and maintain. Before you know it, you’re spending more time tinkering with the automation than it's actually saving you, and it’s still not as accurate as you need it to be.

Using AI for deeper CSAT analysis

Instead of wrestling with brittle workflows or relying on a simple score, you can use AI to analyze the actual words your customer uses and apply much more meaningful tags automatically. This is where a tool like eesel AI can help. It connects directly to helpdesks like Intercom without making you switch your existing tools.

Go beyond scores with smarter analysis

While Intercom's workflow can tag a conversation as "negative," eesel AI’s AI Triage product can actually read the customer’s comment and understand the real reason for their frustration. It can then apply far more useful tags on its own, like:

• "bug-report"

• "feature-request"

• "ui-confusion"

• "slow-response-time"

• "billing-issue"

This gives you an immediate, at-a-glance dashboard of the root causes behind your CSAT scores, without a single person having to manually read through tickets.

A tool that learns from your team

Getting this set up doesn't take months of implementation or a team of developers. With eesel AI, you can be up and running in minutes. You just connect your Intercom helpdesk with a click, and the AI starts learning from your past tickets to understand your product and common customer issues.

And unlike tools that lock you into rigid, predefined rules, eesel AI gives you a flexible workflow engine. You can use its simulation mode to test the AI on thousands of your historical tickets before it ever touches a live customer conversation. This lets you see exactly how it will perform and tweak its behavior with confidence, completely removing the guesswork.

Feature	Standard Intercom Workflow	eesel AI Triage
Tagging Basis	CSAT score (Good/Bad)	Content analysis of customer feedback
Insight Level	Basic (What happened)	Deep (Why it happened)
Setup Time	Minutes to hours	Minutes
Flexibility	Limited by the rule builder	Fully customizable actions & prompts
Testing	Live testing only	Risk-free simulation on past tickets

Creating actionable insights with Intercom workflows

Setting up Intercom workflows to tag conversations for CSAT analysis is a great starting point for any support team that wants to get more organized. It automates a simple but valuable task and helps you quickly sort the good feedback from the bad.

But to really understand your customers and make their experience better, you need to go beyond the score. By using AI to analyze the substance of their feedback, you can stop wasting time on manual analysis and start focusing on making smart, data-driven improvements that actually boost customer satisfaction.

Ready to uncover the real story hiding in your CSAT scores?

See how eesel AI can automatically analyze, tag, and report on your customer feedback. You can start a free trial today.