Skip to main content
Set Fin AI Agent live over chat

How to set up and go live with Fin AI Agent over chat with control and flexibility.

Beth-Ann Sher avatar
Written by Beth-Ann Sher
Updated yesterday

This article covers how to configure and set Fin AI Agent live over chat with control and flexibility over how Fin behaves. You'll need to be using Fin on trial or paid.


Step 1: Install the Messenger

To set Fin live over chat you must install the Messenger so that customers can contact you via chat support.

Step 2: Identify the conversations you want Fin to handle

You'll need to identify the audience or support topics you want Fin to handle over chat and set a success metric (e.g. resolution rate, CSAT). For example, this could include specific high-volume topics or a sub-set of customers.

Step 3: Prepare Fin's knowledge

Go to AI & Automation > Fin AI Agent > Setup and ensure you've enabled content which Fin can use to support chat queries. Update your content to cover your frequently asked chat queries. You can add public articles, Custom Answers, PDFs, snippets, URLs, or content from previous conversations.

You can also add data & actions for Fin to retrieve information from your external systems and apps to deliver personalized answers (or set this up later).

Step 4: Run a small-scale test

Preview Fin over chat

Here's how to preview Fin over chat before setting it live:

  1. Go to AI & Automation > Fin AI Agent > Overview and open the Fin over chat tab.

  2. Use the interactive Messenger preview on the right to start typing questions you think Fin should be able to answer from the content you've enabled.

  3. Once you've previewed the experience, you can finish configuring and testing Fin (see below).

Note:

  • You won't be charged for conversations generated through this preview.

  • Previewing Fin creates a test user called "Preview User" in your inbox.

  • The preview does not respect content audience targeting and will use any of the content available to Fin.

  • The preview will follow any handover setup you've configured if Fin cannot resolve the conversation (i.e. routing to a team inbox).

  • The preview will not send a CSAT survey at the end of your conversation.

Learn more about testing and rolling out Fin to your customers.

Test Fin over chat

Now you can set Fin over chat live to a small test or internal audience (e.g. yourself and teammates) and monitor Fin’s replies for quality and accuracy.

  1. Go to AI & Automation > Fin AI Agent > Overview and open the Fin over chat tab.

  2. Under the "Simple setup", open Who can see Fin and add an audience rule for Email contains [yourcompanydomain.com] - this will enable you and your teammates to send some test questions for Fin to answer.

  3. You can also set up the following:

    • Handover experience from Fin to a teammate.

    • When to auto-close conversations.

    • When a conversation rating (CSAT) is requested.

  4. Select Set Fin live.

  5. Open the Messenger (installed on your website or app) and send a question to see Fin's response.

Note: If you test Fin in your live environment, you will be billed per resolution. However, if you or your teammate has asked to "Get more help" or speak to your team, the conversation is considered "unresolved" and you would not be charged for this.

Review Fin’s responses

Create a set of control questions to share with your team and send via your Messenger, as if you were a customer requesting support.

Then ask your colleagues to note whether Fin provided the best response or not. This can also be a great way to identify gaps in your support content.

If Fin didn’t answer correctly, review the sources used to provide the answer (which are linked in Fin’s response) and update the content so that Fin can provide a better answer next time.

Step 5: Optimize and rollout

Expand the audience Fin is replying to, so that Fin is replying to some real customer conversations. Review the results and customer feedback then expand Fin to handle more of your volume or topics.

We recommend rolling Fin AI Agent out in stages, such as starting with lower-tiered customers who would benefit the most from self-service. This approach allows you to supervise Fin closely and feel confident with how Fin works over chat. See section below for step-by-step guidance.


How to configure your Fin over chat settings for rollout

Simple setup (without workflows)

Go to AI & Automation > Fin AI Agent > Overview and open the Fin over chat tab. At the top of this page, you'll see "Simple setup" where you can:

  • Determine who will see Fin in chat conversations.

  • Configure the handover experience from Fin to a teammate.

  • Choose when to auto-close conversations.

  • Ask for a conversation rating (CSAT) after Fin's conversation.

Note: The "Simple setup" configured from Fin AI Agent takes priority and will trigger before live Workflows. Please keep this in mind if using both "Simple setup" and customer-facing workflows.

Who will see Fin

First, decide which of your customers can interact with Fin AI Agent. You can allow Fin to provide answers to your users, leads, and/or visitors.

Click + Add audience rule and use data attributes in Intercom to target your audience further. For example, Fin could be enabled for conversations with users on your free plan.

Now select which chat channels you'd like Fin to be available on.

Functionality may vary by channel. On Web, Fin uses articles from the Help Center connected to your Messenger brand style. On mobile apps and other channels, Fin uses articles from all live Help Centers.

You can also use channel attributes in the audience rules to configure channel-specific actions and content.

Then click Save and move to the next step.

Let Fin introduce itself

Now configure Fin's introduction message to customers who reach out through chat. This helps your customers better understand what Fin can do and how to interact with it.

You can add/remove and edit the message text fields.

Then click Save and move to the next step.

If Fin cannot resolve the conversation

Specify what action should be taken If Fin cannot resolve the conversation. You can choose to pass to team and assign to a particular team or teammate.

Or, choose to close conversation and offer alternative ways for the customer to get help. For example, provide a contact number or link to another resource.

Then click Save and move to the next step.

Ask for more information before handover

To gather more information from a customer before Fin hands over to a teammate, simply toggle on "Collect more information when a customer asks to speak to the team".

Fin will encourage the user to share more information. The exact message it sends will be something along the lines of "Sure! While I connect you, could you provide us with more details about your issue to help us find an answer faster?" This will be adapted to the context of the conversation and your workspace’s Fin personality settings.

By prompting the customer for more context before Fin hands over to a teammate, there is another opportunity to get more and/or better information than presented the first time around. The benefit to this is two-fold:

  1. Gives Fin another chance to answer: often customers don’t give the AI Agent a chance, asking to speak to a human before asking Fin a question.

  2. If context is provided, even if Fin can’t provide an answer, the teammate will spend less time gathering context and will be able to troubleshoot faster.

We've tested this feature extensively at Intercom and observed an increase in answer rate, confirmed resolutions, and CSAT.

Important: If the user does not reply to Fin's prompt for more information, or if Fin is unable to find an answer to a new user query, the conversation will be automatically routed to the team (and thus will not count as a resolution).

If Fin does find an answer to a new user query, then it will only route them to the team if the user clicks "I still need help 👤". If they click "That helped 👍", the conversation will be regarded as a confirmed resolution, and if they abandon, as an assumed resolution.

Then click Save and move to the next step.

Auto-close abandoned Fin conversations

Choose when to auto-close conversations by specifying a length of time Fin should wait before closing conversations.

Select if you want Fin to automatically close the conversation after customer becomes inactive:

  • If Fin has answered the question, and/or

  • If Fin was not able to answer the question or the customer left before asking any questions.

You can also Customize what Fin says when closing the conversation to make sure the customer got what they needed and remind them they can still speak to your support team, even if the conversation is being closed.

Then click Save and move to the next step.

Ask for conversation rating (CSAT)

At the end of a customer's interaction with Fin AI Agent you can ask them to give a conversation rating by toggling these settings on:

  • 👍 Send CSAT when the customer clicks "That helped" button or replies with a positive message

  • ⏳ Send CSAT if the customer becomes inactive after Fin showed an answer

Then click Save to finish setting up Fin over chat.

Review and set live

Once you're done, click Set Fin live 🎉

You'll need billing permissions to set Fin live if the T&Cs haven't been accepted yet. Accepting these terms is mandatory to be able to use Fin AI Agent.

Advanced setup through Workflows

Go to AI & Automation > Fin AI Agent > Overview and open the Fin over chat tab. At the bottom of this page, open "Advanced setup through Workflows" and select View all workflows.

If you used the "Simple setup" to do some initial testing, we advise that you select Pause Fin to prevent this from triggering before your live Workflows.

If you have an existing workflow set up specifically for inbound conversations through the Messenger or other chat channels:

  • You can simply add Fin to the workflow, where appropriate.

  • When you add Fin to your existing workflows, you'll find these under AI & Automation > Workflows if you filter by "Type is Using Fin".

If you don't have a workflow set up specifically for inbound conversations through the Messenger or other chat channels:

  • We recommend you create one using the "When customer opens a new conversation in the Messenger" or "When customer sends their first message" trigger.

  • You can ensure a workflow is only enabled for chat channels in the trigger settings.

Have Fin respond to a specific audience

By using branching, you can create different Fin experiences for different audiences.

For example:

  • For paying customers, you might want to hand over to a teammate when Fin can't answer but for non-paying customers, you might want to direct them to your Help Center.

  • See additional ways and examples of how you can target specific audiences or customers.

Add as many branches as you need to create different experiences for your different customers.

While you can also determine audience rules within the workflow trigger settings, we recommend only doing this if you have an additional workflow to cover everyone who doesn't match those rules.

Have Fin respond to a specific topic

You can also use branching to create different Fin experiences and paths based on your different conversation topics.

  1. Click Add step to start creating your branching logic.

  2. Then select Branches under "Proceed to another path".

  3. Once you’ve added a branch, click on Missing condition to open up your branch settings.

  4. Click + Add condition to set different conditions, e.g. if you want to route conversations based on a keyword in the message, select "Message Content" as a condition.

  5. You can repeat this step to add multiple branches per keyword.

  6. Once you’ve defined your branches, click on the arrow to the right to determine the path specific to that branch, e.g. assign to a teammate or let Fin answer.

For example, you can use a "Message Content" filter to route conversations which contain the keyword "billing" directly to your Finance team.

With AI category detection [closed beta], Fin can categorize conversations by topic and use these AI categories as conditions in workflows to automatically route conversations to the right team, ensuring customer requests are handled efficiently.

Add Fin to your workflow

Once you’ve determined the paths for the different audiences and/or topics, you can then select where you want Fin involved. Add Fin to your workflow by creating a path and selecting Let Fin answer.

Customize Fin's behavior in your workflow

To customize Fin's behavior in a workflow, click on the Let Fin answer step in the workflow path. Here, you can:

  • Set expectations for human support.

  • Control the answer type.

  • Configure the handover to teammate experience.

Test your Fin over chat workflow

If you want to test a Fin over chat workflow, select the Preview button at the top of the Workflows builder. This will give you an interactive preview of the Messenger experience when talking to Fin.

Best practices for chat workflows

  • The "Simple setup" from Fin AI Agent takes priority and will trigger before live workflows. Please keep this in mind if using both "Simple setup" and "Advanced setup through Workflows".

  • We don't recommend having multiple workflows per topic. We recommend you have one workflow for your inbound chat, with each conversation topic defined using branches.

  • When creating a workflow, you should always set a path for "Else". In the event a customer doesn’t fall under a specific topic or audience, this ensures there is a fallback option to support them.


Messenger CTA

When you've set Fin AI Agent live for customers, the call-to-action (CTA) button which allows customers to start a conversation in your Messenger may change.

This depends on whether you've configured your Messenger settings to:

  • Set bot expectations, or

  • Match teammate expectations.


💡Tip

Need more help? Get support from our Community Forum
Find answers and get help from Intercom Support and Community Experts


Did this answer your question?