Intercom Fin 2: What’s new, features, and setup guide

If you’ve worked in customer support lately, you know the drill. Tickets pile up faster than coffee orders on a Monday morning, and everyone wants answers yesterday.

That’s where an AI copilot comes in. It’s a virtual assistant that never sleeps, never sighs, and claims a 31% increase in agent efficiency. It cannot make you a latte, but it does promise to make life a little easier. With the release of Fin 2, Intercom has leveled up its assistant by improving context awareness, response quality, and how it fits into your daily support workflows.

It pulls from your help center, past conversations, and internal documents to suggest accurate, personalized replies. It is like giving your agents a second brain, just without the burnout.

In this guide, we will explore how Fin AI works, where it performs well, and where it may still fall short. We will also look at other tools like eesel AI that offer more flexibility and work across a wider range of platforms. Whether you are exploring AI for the first time or looking to upgrade your current setup, this is a good place to start.

How does Intercom’s Fin 2 AI copilot work?

Fin 2 is Intercom’s latest version of its AI support agent. It’s built to help teams respond to customer questions across multiple channels, using content pulled from help centers, internal docs, and other sources.

At the core of Fin 2 is the Fin AI Engine, which aims to improve how customer queries are understood and answered. It uses a combination of language models, source checking, and your company’s support content to suggest accurate replies. According to Intercom, this setup helps Fin 2 balance response quality with fewer incorrect answers.

Fin 2 works across channels like chat, email, and messaging apps. You can manage how it responds, what content it draws from, and how it interacts with your human team. Features like tone presets, response length options, and category detection give you more control over how the assistant fits into your support process.

A look at what powers Fin 2

Fin 2 can also pull information from multiple sources to answer more complex questions, and it supports real-time translation in over 45 languages. If a question falls outside its capabilities, it can hand off the conversation to a human agent based on the rules you’ve set.

What sets Fin AI apart from traditional support tools is its sophisticated information synthesis capabilities. Unlike basic chatbots that pull from single knowledge bases, Fin AI intelligently combines and cross-references multiple sources while maintaining accuracy through built-in source verification. Each suggested response includes links to source material, allowing agents to quickly validate information before sending it to customers.

What’s new in Fin 2

Fin 2 builds on the foundation of Intercom’s original AI agent but introduces a long list of improvements. Some features are available now, while others are still rolling out. These updates fall into four main categories: knowledge, behavior, actions, and insights.

Feature	What it does
Knowledge Hub	Centralizes all content Fin learns from. Supports help center articles, internal docs, PDFs, and webpages. Updates automatically.
Multi-source answers	Combines multiple sources to generate more complete, context-rich responses.
Content targeting	Allows different responses based on customer plan, location, or brand.
Tone and length controls	Adjusts Fin’s tone (e.g. professional, friendly, humorous) and response length (short or detailed).
Real-time translation	Supports 45+ languages, even if content isn’t localized.
Tasks and custom actions	Enables Fin to perform actions like checking orders or updating subscriptions when connected to external systems.
Workflows and category detection	Lets you create rules for how Fin handles or routes certain topics, like refunds or bug reports.
Expanded channels	Works across chat, email, WhatsApp, SMS, Facebook, and Instagram.
Holistic reporting	Provides reporting across both human and AI support, including unresolved questions and AI performance insights.

How Fin 2 fits into your workflow

You can adjust how Fin 2 shows up for different customer segments, regions, or channels. It’s also possible to connect it to certain external systems, which allows it to take limited actions on behalf of customers, such as checking order status or updating an account.

That said, Fin 2 is still best suited for teams already using Intercom as their main support platform. If your workflows span multiple systems or require a broader set of integrations, tools like eesel AI may be worth considering. These tools offer more flexibility when it comes to connecting content across platforms like Google Drive, Slack, Shopify, or Confluence.

Setting up Fin 2: What’s new and what to prepare

If you’ve worked with Fin 1, you’ll notice that Fin 2 doesn’t just add new features — it reshapes how setup and training work. Many core steps are still familiar, like connecting your knowledge base and defining handoff rules, but Fin 2 gives you more control, more customization, and new tools to manage the entire process.

Key differences in the setup process

Fin 1	Fin 2
Relied heavily on help center content	Uses a centralized Knowledge Hub that supports help center articles, internal docs, PDFs, webpages, and more
Single-source answers	Fin 2 can generate responses from multiple content sources at once
Limited tone control	You can now select from predefined tone presets (e.g. professional, friendly, humorous) and adjust response length
Manual content targeting	You can target content by customer type, location, or brand directly in the interface
Basic configuration only	Includes category detection, workflows, and task automation to personalize how Fin handles different topics
Chat-focused	Works across chat, email, WhatsApp, SMS, Facebook, and Instagram right from setup

Steps to get started with Fin 2

1. Prepare your content in the knowledge hub

Before turning Fin 2 on, make sure your content is accurate, well-organized, and up to date. The Knowledge Hub is where you’ll manage everything Fin learns from, including help center articles, internal docs, and other file types. Unlike before, content updates here are automatic — you do not need to retrain Fin every time something changes.

However, Fin’s accuracy still depends heavily on how well your documentation is written. If the content is sparse, scattered, or inconsistent across tools, expect more misfires or fallback responses.

2. Set tone and response style

ou can define how Fin responds by selecting tone presets like “professional” or “friendly” and choosing between short or long replies.
While this makes it easy to get started, the options are limited. You can’t give detailed prompts or granular language instructions. If your team supports multiple brands or needs more nuanced communication styles, this could feel restrictive.

3. Define rules with guidance and workflows

Fin supports natural language coaching and basic workflows for routing topics, like escalations or handoffs for specific keywords.
The tradeoff is that these workflows are relatively simple. More advanced branching logic or scenario-based handling may require workarounds, especially for multi-department setups or niche use cases.

4. Connect to channels and external systems

Fin works across chat, email, WhatsApp, SMS, and social platforms like Facebook and Instagram. It also connects to certain systems to perform tasks like order checks or account updates.
That said, connectivity is limited to integrations already supported by Intercom. If your workflows involve external platforms or rely on content stored in tools like Google Drive, Confluence, or Slack, you may hit integration barriers. eesel AI is built to work across over 100 tools out of the box, making it easier to automate complex workflows without custom development.

5. Test before going live

Intercom provides tools to test Fin’s replies before launch. You can preview its responses, simulate conversations, and monitor how it handles specific types of questions.
Still, these simulations don’t fully replicate every edge case. Teams managing sensitive workflows or high support volumes may want to layer in extra review steps or limited rollouts to avoid surprises.

6. Launch gradually and monitor performance

Once live, track Fin’s resolution rate, accuracy, and areas where customers still request help. Use Intercom’s reporting tools to review unresolved questions, identify weak spots in your content, and continuously fine-tune how Fin works.

Comparing AI customer support assistants

Not every support team needs the same thing from an AI assistant. Fin 2 introduces a more capable and configurable version of Intercom’s original AI agent, with expanded support for new content types, external systems, tone settings, and additional communication channels. It works best for teams already operating within the Intercom ecosystem. Setup is straightforward when your content, workflows, and automation rules are already housed inside Intercom.

However, Fin 2 may be less flexible for teams that rely on a wider mix of platforms. If your support content lives in places like Google Drive or Confluence, or your workflows span across multiple systems, Fin may not fully adapt to how your team works. While some integrations are supported, the product is still most effective when used within Intercom’s own environment.

If your setup includes several external tools, it may be worth exploring how alternatives like eesel AI compare in terms of integration, customization, and overall fit.

Fin 2: Where it’s limited

Fin 2 is built for teams fully committed to Intercom. While it performs well inside that ecosystem, it comes with a few tradeoffs that might limit flexibility for broader setups.

Integrations

Fin supports select platforms, but most workflows depend on centralizing content in Intercom. If your docs live in tools like Google Drive or Confluence, syncing them requires extra steps. eesel AI connects to over one hundred platforms and syncs content directly, without needing to migrate everything manually.

Customization

Fin offers tone presets and response length controls, but customization ends there. You cannot easily define unique greetings, brand-specific language, or emoji use. eesel AI allows full prompt control, so your bot speaks exactly how you want across every channel.

Training

Fin auto-learns from past tickets, but it lacks filtering or editing. If those conversations are outdated or inconsistent, Fin can repeat poor answers. eesel AI turns tickets into editable articles, letting you decide what to include and update.

Pricing

At ninety-nine cents per resolution, Fin’s costs can stack up quickly. Extra fees for channel usage add more unpredictability. eesel AI’s interaction-based pricing avoids per-resolution fees and keeps monthly billing simpler and easier to manage.

Where eesel AI beats Fin 2

eesel AI offers more flexibility for teams using tools beyond Intercom. It connects with over 100 platforms, including Google Drive, Slack, and Confluence, making it easier to work with your existing content and systems.

It supports over 100 languages, offering broader international reach than Fin’s 45. You also get deeper customization, including tone control, department-specific workflows, and advanced API actions.

eesel’s Team Plan costs $239 per month and includes up to 3 AI bots, 1,000 interactions, and integrations with tools like your help desk and Slack. Pricing is based on usage rather than per resolution, which may be more predictable for teams handling larger volumes or sharing workloads across agents.

Let AI do the heavy lifting

At the end of the day, there is no single “best” AI solution for every support team. It all comes down to finding the right fit for your tools, your workflows, and your customers’ needs. If you are already invested in the Intercom ecosystem, Fin 2 can be a natural choice. If your team relies on multiple platforms or needs more flexibility and control, eesel AI may be the better option.

Curious to see how it all works? Start your free trial, book a personalized demo, or email us anytime at hi@eesel.app.