At Intercom, our approach to leveraging AI and automation is centered around three interconnected features:

This guide will help you to use these features across your support operations and successfully implement AI-first customer service.

We’ve broken it up into the following sections:

By taking advantage of workflow automations you can scale your support offering while making interactions with customers much more effective, and drive significant improvements across core metrics like first response time and CSAT. Once you’ve mastered Workflows, you’ll be able to take care of all your automations in Intercom, including implementing an AI agent to handle support queries (which we’ll cover next!)

We recommend watching these four short videos which explain the basics of how to use workflows (8mins). This will help set you up for success with building automations in Intercom, and once you have this foundational knowledge, you can dive deeper using the resources below:

To meet your expected reply times and build customer trust, you should set Service Level Agreements (SLAs), and allow your team to stay focused on active issues by automatically managing inactive conversations. Of course, the true test of Customer Satisfaction (CSAT) is asking them, so you can also follow up any conversation with a CSAT survey and gain valuable feedback from your customers.

Now that you’ve got the hang of workflows, you can simply add Fin to instantly resolve up to 50% of your support queries. Using Fin in workflows improves your customer experience by reducing wait times and boosting CSAT scores, and frees up your team to focus on more important things. All Fin needs is your support content - the better your content, the better Fin performs!

If you need a bit more control and diversity with the support Fin provides, you can customize the experience with Custom Answers and enable multilingual Fin to support customers in 45 different languages. Using Custom Answers also allows you to embed images, videos and apps, or retrieve external data and take actions during a conversation.

Cater for a global audience and allow Fin to provide support in 45 languages by enabling different language versions of your content. This can also provide a multilingual help center for your customers to self-serve.

Custom Answers are always prioritized over AI answers from your support content, and you can apply conditional logic to target specific customers with specific follow-up actions. For example, an answer might be different if a customer is on a paid plan vs. a free plan. You can also add Custom Actions, Custom Objects and/or Custom Apps to fetch external data and take actions like finding a customer’s order and processing a refund.

After adding a new AI agent to your support team, it’s important to check in on their performance (like any other teammate!) You can measure Fin’s impact on your support volume, CSAT, and first response times, and see which topics need more knowledge coverage to keep driving resolutions up and saving hours of support time.

When adding Fin to your workflows, turn on ask for a conversation rating at the end of a customer’s interaction with Fin. Then you can use the built-in Fin CSAT report to find the exact rating per conversation, and filter by neutral/negative CSAT to find conversations where Fin’s interaction could be improved either through the workflow where Fin was used, or in the knowledge Fin has access to.

The most impactful ways to continuously improve Fin’s knowledge and resolution rate are: collaborating with customer-facing teams to identify the high volume questions they receive and ensure the answers are captured in the Knowledge Hub for Fin to use; asking subject matter experts to review your content and make sure it’s still up-to-date and relevant; and ensuring your content format and structure is optimized for AI.

With workflows and Fin working nicely in the background, these inbox AI features are designed to help your teammates directly. They can speed up writing replies and reduce cognitive load by automating tasks which often require extra time or skills. This means your support team works more productively and to the same high standard whether they’re a tenured agent or new to the team, ensuring customers receive exceptional service every time.

So now it’s easy to compose replies, but you still need to provide the right answer and Support teams lose valuable time hunting for answers across multiple tabs and sources. With Fin AI Copilot, your team gets instant, expert answers which can be sourced from your recent conversation history, internal articles, public articles, and external content, all directly within the inbox. This allows you to increase your team’s efficiency up to 31% so agents can spend more time focusing on what really matters: building deeper, lasting customer relationships.

First, you’ll need to give Fin AI Copilot access to your internal support content.

Then decide which teammates should have access to Fin AI Copilot and what level of usage they need. Included usage (up to 10 conversations per month) is available to all teammates by default, but you can also choose to give some teammates unlimited usage, or turn it off for them completely.

Fin AI Copilot is easily accessed from the right sidebar in the inbox and very intuitive to use, but to encourage strong team adoption and get the best results when asking Fin a question, you should share these resources with your teammates.

If you're excited to harness the power of AI-first customer service, check out our Academy courses which guide you through getting started with the Knowledge Hub, Fin AI Agent, and Fin AI Copilot.