Fin AI

What is Fin AI?

Fin AI is Intercom’s powerful AI assistant that streamlines support operations by automating conversations and tasks. It combines natural language processing, machine learning, and deep integrations with your existing systems to provide intelligent, personalized support at scale.

This support tool empowers your team to work smarter, not harder, by handling routine inquiries, triaging complex issues, and augmenting human agents with contextual insights and recommendations. With Fin AI, you can elevate the customer experience while boosting team productivity and efficiency.

To get started, simply connect your data sources, and it will learn from your knowledge base, past conversations, and business logic. From there, it continuously improves, adapting to your evolving needs and customer conversations.

Core Features

• Intelligent Conversation Handling: Fin AI can comprehend and respond to customer inquiries in natural language, providing accurate answers or routing complex issues to the appropriate support agent.
• Task Automation: With deep integrations into your systems, it can automate various tasks, such as updating order statuses, processing refunds, or retrieving customer data, saving your team valuable time and effort.
• Contextual Recommendations: By understanding the context of each conversation, it can provide agents with relevant information, knowledge base articles, and next-best-action recommendations, empowering them to resolve issues more efficiently.
• Multi-Channel Support: It operates seamlessly across multiple channels, including email, chat, social media, and messaging apps, ensuring a consistent and cohesive customer experience.
• Continuous Learning: Fin AI’s machine learning capabilities allow it to continuously improve its understanding of your business, products, and customers, providing more accurate and relevant responses over time.

Pros & Cons

Pros	Cons
Efficiency gains: Significant time and cost savings by automating routine support tasks	Setup investment: Initial setup and configuration may require dedicated resources
Enhanced satisfaction: Improved customer satisfaction through faster response times and personalized support	Accuracy challenges: Potential for inaccurate responses or misunderstandings, especially in complex scenarios
Smooth integration: Seamless integration with existing systems and knowledge bases	Handling limitations: Limitations in handling unique or nuanced customer queries
Adaptive flexibility: Continuously evolving and adapting to your business needs	Data reliance: Reliance on high-quality training data and knowledge bases
Scalable capacity: Scalable and capable of handling high conversation volumes	Maintenance demands: Ongoing maintenance and monitoring required to ensure optimal performance

Pricing

Category: Paid

Fin AI Agent is priced at $0.99 per resolution (United States Dollars) and is available on every Intercom Core plan at no additional cost. This means that each time Fin successfully resolves a customer inquiry, you’ll be charged $0.99.

You’ll only incur charges when you actively use Fin AI Agent. If you choose not to use it, there are no additional costs. You can manage your usage by setting reminders and limits to suit your needs.

How to Set Up Fin AI

Setting up Fin AI is a straightforward process that involves connecting your data sources and configuring the AI assistant to your specific needs. Intercom offers two primary methods for setting up depending on your plan.

Workflows is the recommended method if you have a paid Intercom plan. It allows you to fully customize Fin AI’s behavior and interactions.

Setting Up Fin AI with Workflows (Paid Plans)

Workflows provide an extremely flexible and customizable way to configure Fin AI’s behavior, visibility, and integration with your support processes. Let’s dive into the step-by-step process for getting up and running using Workflows.

1. Accessing Workflows

The first step is to make sure you have access to the Workflows feature based on your current plan.

Navigate to the Fin AI Agent section, then click on “Fin Overview“. Here you’ll see the “Set up and go live” tab – open this.

2. Choose Your Setup Method

You now have two options for setting up Fin AI with Workflows:

1. Quick Setup: This allows you to rapidly deploy the AI by configuring just a couple of high-level settings. While not as customizable as building your own workflow, it’s a fast way to get started.
2. Tailor Your Own Workflows: This provides much more flexibility to customize every aspect of the AI’s behavior, audience targeting, handover processes, and more. Let’s cover this option in depth:

3. Starting a New Workflow

If building your own tailored workflow, you have two choices for how to get started:

1. Start with a Template: There are pre-built workflow templates designed specifically for Fin AI that can streamline the setup process. These templates maximize Fin’s visibility while still providing a good handover experience to your human teammates when needed.
2. Add Fin to an Existing Workflow: Alternatively, you can add the Fin AI component to one of your existing workflows that you’ve already built out. This is a good choice if you have an established workflow you want to leverage.

If you choose to “Add Fin to an Existing Workflow,” you’ll be directed to your Workflows page. There, you can open and modify one of your preexisting workflows to incorporate Fin AI Agent.

The best workflow triggers to use with Fin AI are:

• “Customer opens a new conversation in the Messenger” – Use this to have Fin AI introduce itself immediately in new conversations
• “Customer sends their first message” – Use this to have Fin AI respond to the initial message across all channels like WhatsApp, SMS, etc.

To integrate Fin AI Agent into your Workflow, you’ll need to create a path and then choose “Let Fin answer.”

4. Configure Fin AI Behavior

With your new or existing workflow opened, there are several aspects of Fin AI’s behavior that can be customized:

1. Choose Channels: If using the “Customer sends any message” trigger, you can specify which channels Fin AI should operate in – e.g. Messenger, email, WhatsApp etc.
2. Set Expectation for Human Support: By default, Fin AI provides a “Talk to a person” button to allow handoffs to your human teammates. However, you can disable this if you don’t want to set expectations for human support.
3. Auto-Close Parameters: Define when Fin AI should automatically close inactive conversations after a period of time. You can set custom timeframes for this across different scenarios like:
   • User hasn’t responded
   • User responded with “That didn’t help”
   • User didn’t respond after being handed off to a teammate
4. Customize Closing Messages: When auto-closing inactive conversations, you can customize the language Fin AI uses in its closing message to provide resources or encourage the user to re-open if needed.
5. Conversation Ratings: You can optionally enable prompts for users to rate their conversation experience with Fin AI after it has attempted to provide a resolution.
6. Content Sources: If you’ve configured custom answer articles, you can specify whether Fin AI should only suggest custom answers, or blend them with its general AI-generated responses.

5. Configure the Handover Experience

A key aspect of the workflow is defining what should happen when Fin AI cannot fully resolve the user’s question, and they request to “Talk to a person”. Here are the options:

1. Handover to Teammate: The most common option is to trigger a handover to one of your human support teammates or teams at this point. You can fully customize this process with additional steps:
   • Setting expectations for human reply times
   • Collecting additional details from the user
   • Automatically routing to the right support team/person
   • Using conditional branching logic based on user attributes
2. Pass to Reusable Workflow: Another option is to pass the conversation to an existing reusable workflow you’ve built. This is useful if you have a common triage or data collection process.
3. Configure with Branching: Workflows allow branching logic based on audience attributes. This enables creating different experiences for different user segments.

For example, you may want to route paid customers to a teammate right away when Fin AI can’t help, while directing non-paid users to your knowledge base.

6. Preview and Deploy

Once you’ve completed building out your workflow, you’re ready for some final testing before deployment:

1. Preview Mode: Use the Preview button in the workflow builder to experience an interactive test conversation with Fin AI based on the current configuration. Fix any issues that arise.
2. Set Live: When you’re fully satisfied, click the “Set Live” button to deploy the workflow and make Fin AI operational according to your configurations! You’ll need billing permissions to set it live.

7. Post-Launch Experience

After setting the Fin AI workflow live, you’ll have access to some additional capabilities:

• The Fin Inbox for viewing/managing conversations
• Reporting on conversation stats and resolutions
• Fin AI optimization tools to identify areas for improvement
• Cost monitoring and billing information

Setting Up Fin AI Without Workflows (Starter or Essential Plans)

If you’re on a Starter or Essential plan without access to Workflows, there is an alternative setup path for configuring Fin AI. Here are the steps:

Prerequisites

Before you begin, make sure you have:

1. Access to your Intercom account with the appropriate permissions.
2. Support content (e.g., Help Center articles, API docs, PDF guides) that you want Fin AI Agent to learn from.

1. Add Content

1. Log into your Intercom account and navigate to the Fin AI Agent section.
2. Click on the “Set up Fin” button on the Overview page.
3. On the “Add content” tab, click “Add content source” to connect your data sources (Help Center articles, API docs, PDF guides, etc.).
4. Alternatively, you can drag and drop files or click to upload PDFs, web pages, Markdown files, etc.
5. Preview the content on this page to ensure it has imported correctly.

2. Set up and Go Live

1. Navigate to the “Set up and go live” tab.
2. On the right side, you’ll see an interactive preview window where you can test asking Fin questions based on your uploaded content.

3. Configure Audience and Channels

1. In the “Who will see Fin” section, select the target Audience(s) you want Fin to be available for.
2. Choose the Channels (e.g., Messenger, email) where Fin should operate.
3. Click “Save” to apply the changes.

4. Customize Introduction

1. In the “Let Fin introduce itself” section, check the language(s) you want Fin to operate in.
2. For each language, customize the introduction message Fin will use.
3. You can add multiple variations of the introduction message that will rotate.
4. Click “Save” to apply the changes.

5. Configure Unresolved Conversations

1. In the “If Fin cannot resolve the conversation” section, choose to either Close the conversation or Pass to a team if Fin can’t help.
2. If passing to a team, select which Team or individual to assign to.
3. Set the time period for auto-closing inactive conversations.
4. Customize the closing message Fin will use.
5. Click “Save” to apply the changes.

6. Enable CSAT Survey (Optional)

1. In the “Ask for conversation rating when Fin finishes” section, toggle on to enable the CSAT survey.
2. Set the number of minutes before the rating can’t be changed.
3. Click “Save” to apply the changes.

7. Test and Go Live

1. Use the interactive preview pane to thoroughly test asking Fin questions from your content.
2. Once you’re satisfied with the setup, scroll to the bottom and click the “Set Fin Live” button.

Upgrade/Downgrade Plans

If you upgrade to a higher plan with Workflows, your Fin setup will automatically migrate to Workflows, and a new Workflow will be created based on your prior configuration.

If you downgrade to the Starter/Essential plan, any Fin Workflows will be paused, and you’ll need to go back to the Overview > “Set up Fin” to re-activate Fin AI Agent.

No matter which setup method you use, Fin AI provides powerful AI assistance capabilities along with robust customization options to mold it to your team’s specific needs and processes. Take full advantage of the preview testing capabilities to optimize the experience before launch.

With thoughtful configuration, Fin AI can drastically increase your support team’s efficiency while providing an exceptional customer experience.

Walk through each of the options methodically, as the setup choices you make will significantly impact the tool’s effectiveness. Don’t hesitate to iterate and refine the configuration over time as you learn how users respond.

Managing and Optimizing Fin AI

To get the most out of Fin AI, proactively manage and optimize its performance:

1. Training and Continuous Learning: Regularly provide Fin AI with new training data, such as updated knowledge base articles, product information, or customer conversations.
2. Performance Monitoring: Use Intercom’s reporting and analytics tools to monitor Fin AI’s performance metrics, like success rates, resolution times, and customer satisfaction scores. Identify areas for improvement.
3. Agent Feedback Loop: Encourage your support agents to provide feedback on Fin AI’s responses and recommendations. Use this input to refine Fin AI’s capabilities.
4. Conversation Review: Periodically review Fin AI’s handled conversations to identify inaccuracies or instances where it struggled to understand the customer’s intent. Use these insights to improve Fin AI’s training data and conversation flows.
5. Customization and Personalization: Tailor Fin AI’s personality, tone, and language to align with your brand and customer expectations. Customize response templates, conversation flows, and integrations for a seamless, personalized support experience.

These strategies, coupled with Intercom’s robust management tools, empower you to maintain Fin AI’s excellence and elevate your customer experience consistently.

Using PDF Files as Content Sources

Fin AI Agent can leverage PDF files as an additional content source, allowing you to incorporate valuable information from existing documentation and resources. To get started with using PDF files:

1. Access the “Content” section under Fin AI Agent in your Intercom workspace.
2. Click the “New Content” button and select the option to upload a PDF file.
3. Either drag and drop the PDF file into the modal or browse to select the file from your local storage.
4. Once uploaded, the PDF file will be processed, and the text content will be extracted and made available for Fin AI Agent within a few minutes.

The extracted text from the PDF will be treated as a private content source, similar to Snippets. This means that while Fin AI can reference this content when generating responses, customers will not see direct links or citations to the PDF itself.

It’s important to note that Fin AI Agent can only process text-based PDF documents. Any images, encrypted files, or password-protected PDFs will not be supported. Additionally, there is a limit of 100 PDF files per workspace, and each file must have unique content with a maximum size of 45 MB.

Leveraging Conversation Content (Open Beta)

Fin AI Agent’s “Conversation Content” feature allows you to tap into the collective knowledge and experience of your support team. By enabling this feature, you can empower your teammates to suggest excerpts from resolved conversations as potential content sources for Fin AI.

To get started with conversation content:

1. Navigate to the “Overview” section under Fin AI Agent and enable the “Collect conversation content from the inbox” toggle.
2. Select which teammates should be allowed to suggest conversation content, either choosing specific individuals or allowing any teammate to contribute.
3. Configure whether suggested content should be immediately available for Fin AI Agent or set as pending for review.

When enabled, teammates will see a modal after closing or snoozing certain conversations, prompting them to review an AI-generated question-and-answer excerpt from the conversation. They can then choose to suggest this content for Fin AI Agent, edit the suggested content, or decline to suggest it.

Content Suggestion and Review Process

To ensure the quality and relevance of the content used by Fin AI Agent, Intercom provides a robust content suggestion and review process. This process varies depending on your configuration:

1. Immediately Available Content: If you’ve chosen to make suggested content immediately available, any content proposed by teammates will be added to the “Content” page as a new Snippet and immediately used by Fin AI Agent. You can review, edit, or disable these Snippets at any time from the “Content” page.
2. Pending Review Content: Alternatively, you can choose to have suggested content set as “pending” in the “Content Suggestions” page for review. This page allows designated content managers to review each suggestion, preview the content, and either approve or reject it for use by Fin AI Agent.
   • Approved suggestions will be added to the “Content” page as new Snippets, marked as “Used by Fin.”
   • Rejected suggestions will be removed from the “Content Suggestions” page and not used by Fin AI Agent.

To streamline the review process, you can review suggestions one by one or select multiple suggestions for bulk approval or rejection. Additionally, Intercom provides notifications and reminders when there are pending suggestions awaiting review, ensuring timely action and minimizing disruptions to Fin AI Agent’s performance.

Continuous Improvement and Optimization

Intercom recognizes that optimizing Fin AI Agent’s performance is an ongoing process. The “Answer Debugger” feature provides valuable insights to help you identify areas for improvement and take action on outdated or missing content.

From the Inbox, any conversations involving Fin AI Agent will have an “Improve Answer” button next to each AI-generated response.

Clicking this button opens the Answer Debugger in a side view, revealing the following information:

1. Original Customer Question and AI Response: At the top, you’ll see a summary of the customer’s original question and the AI-generated response provided by Fin AI Agent.
2. Relevant Sources Used: The debugger displays a list of the top content sources (up to 10) from your Content library that Fin AI Agent deemed relevant to the customer’s question.
3. Source Interaction Options: For each relevant source displayed, you have the option to open or edit the content directly from the debugger:
   • Intercom Articles: Open the article in a new tab or edit it directly.
   • Snippets: Open the Snippet in the “Content” page or edit it within the debugger view.
   • External Content: Open the external URL source.

By analyzing the sources used and the AI-generated response, you can identify gaps or outdated information in your content. You can then take action by creating new Snippets or updating existing content, ensuring that Fin AI Agent stays aligned with the latest information and best practices.

This continuous improvement cycle, facilitated by the Answer Debugger, allows you to refine Fin AI Agent’s knowledge base and enhance its ability to provide accurate and relevant responses to your customers.

Fin AI Security and Compliance

Intercom understands the importance of data security and compliance, particularly in the realm of customer support and AI-powered technologies. Fin AI is built with robust security measures and adherence to industry standards:

1. Data Encryption: All customer data and conversations handled by Fin AI are encrypted in transit and at rest, ensuring the utmost protection against unauthorized access or data breaches.
2. Secure Infrastructure: Fin AI operates on Intercom’s secure and scalable cloud infrastructure, which is regularly audited and compliant with industry standards such as SOC 2, ISO 27001, and GDPR.
3. Access Controls: Strict access controls and role-based permissions ensure that only authorized personnel can access and manage Fin AI’s configurations and data.
4. Compliance Standards: Fin AI is designed to comply with various industry regulations and standards, including GDPR, CCPA, and PCI-DSS, safeguarding customer privacy and data protection.
5. Regular Audits and Monitoring: Intercom conducts regular security audits, vulnerability assessments, and continuous monitoring to identify and mitigate potential risks or threats.

Intercom ensures that Fin AI operates within a secure and trusted environment, giving you peace of mind when handling customer data and automating support processes.

Fin AI Best Practices and Tips

To maximize the value and effectiveness of Fin AI, Intercom recommends the following best practices and tips:

1. Start with Clear Goals: Define your objectives for implementing Fin AI, such as improving customer satisfaction, reducing response times, or increasing agent productivity. This will help you measure success and optimize Fin AI’s configuration accordingly.
2. Prioritize Training Data: Invest time and effort in curating high-quality training data, including knowledge base articles, product documentation, and sample conversations. The quality of this data directly impacts Fin AI’s performance and accuracy.
3. Encourage Agent Feedback: Actively involve your support agents in the optimization process by gathering their feedback on Fin AI’s performance, areas for improvement, and any recurring customer pain points.
4. Continuously Monitor and Adjust: Regularly monitor Fin AI’s performance metrics and customer feedback, and make adjustments to its configurations, conversation flows, and training data as needed to ensure optimal performance.
5. Leverage Integrations: Take advantage of Fin AI’s integration capabilities by connecting it with your existing systems and data sources, enabling seamless data flow and task automation.
6. Communicate with Customers: Educate your customers about Fin AI’s capabilities and limitations, setting appropriate expectations and providing clear escalation paths for complex or sensitive issues.
7. Foster Collaboration: Encourage cross-functional collaboration between your support, product, and engineering teams to ensure Fin AI’s continuous improvement and alignment with your overall business objectives.

Following these best practices and leveraging Intercom’s resources and support will allow you unlock the full potential of Fin AI and deliver exceptional customer experiences while enhancing your team’s productivity and efficiency.

Conclusion

Intercom’s Fin AI is a powerful and intelligent support automation solution that empowers businesses to deliver exceptional customer experiences while optimizing their support operations. It can streamline routine tasks, augment human agents, and provide personalized, context-aware support at scale.

With its continuous learning and adaptation abilities, Fin AI evolves alongside your business, ensuring that you stay ahead of customer expectations and industry trends. Intercom’s commitment to security, compliance, and best practices further solidifies Fin AI as a trusted and reliable partner in your customer support journey.

Embrace the future of intelligent support automation with Fin AI and unlock new levels of efficiency, personalization, and customer satisfaction.