How to Set Up and Optimize Fin for Zendesk
This guide will help you set up Fin AI Agent to enhance your customer support experience. By following these steps, you'll train Fin with your support content, optimize its responses, and deploy it in Zendesk to efficiently handle customer inquiries.
Executive Summary
The best way to think about an AI agent like Fin is to imagine you’re onboarding a new team member. Just like a new hire, Fin needs:
Access to the right information
Clear guidance on communication style
The ability to learn from experience
By gradually providing Fin with relevant content, refining its behavior, and testing its responses, you ensure it effectively supports your customers. Fin will continuously improve, learning from its interactions and the adjustments you make to its training.
Four-Step Framework for Deploying Fin:
Train: Provide Fin with support content, define its communication style, and set escalation rules.
Test: Validate responses with real customer questions and refine as needed.
Deploy: Set Fin live within Zendesk and define interaction and escalation rules.
Analyze: Monitor performance, optimize responses, and improve Fin’s efficiency.
This guide walks you through each step to help you maximize Fin’s impact on your support operations.
Before You Get Started
How Fin Works in Zendesk
Fin acts as an AI-powered support agent within Zendesk, integrating with:
Knowledge Base: Uses existing help articles for accurate responses.
Messaging & Ticketing: Engages in live chats, handles tickets, and escalates when needed.
Intercom Messenger (if applicable): Supports customers while escalating complex issues to Zendesk agents.
Understanding how Fin fits into your Zendesk environment helps you configure it effectively.
Assemble Your Project Team (optional)
To successfully deploy Fin, form a team with key roles:
Support Leader: Ensures Fin aligns with customer support goals.
Technical Leader: Manages integration and troubleshooting.
Content Owner: Optimizes Fin’s knowledge base for accuracy and relevance.
Define Success Metrics
Track Fin’s effectiveness with:
Resolution Rate & Customer Satisfaction (CSAT): Measures answer quality and customer satisfaction.
Average Handling Time & Total Hours Saved: Evaluates efficiency gains.
Qualitative Factors: Assesses tone, clarity, and answer quality.
Deflection & Involvement Rates: Provides insights, though not always a complete performance picture.
Step 1: Train
Train Fin with Your Support Content
Fin provides accurate responses by leveraging your existing support materials. Ensure Fin has access to high-quality training content:
Sync Content from Zendesk: Connect Fin to your help center.
Upload PDFs or Text Snippets: Supplement with FAQs or policy documents.
Manage Content Updates: Regularly update Fin’s knowledge base.
✅ Sync & Manage Content Sources
✅ Sync public articles with Zendesk
✅ Upload and manage PDF files
✅ Enable or disable content for Fin AI Agent
Train Fin with Guidance
Beyond knowledge, Fin needs structured communication rules to understand how to communicate and when to escalate inline with your companies processes:
Set the Basics: Define Fin’s tone and response length.
Define Communication Style: Guide vocabulary and terminology usage.
Set Up Clarification Steps: Train Fin to ask follow-up questions.
Establish Escalation Rules: Define when and how Fin should escalate issues.
Use Tasks to Automate Multi-Step Processes
Fin can automate workflows beyond simple customer queries:
Create Task-Based Workflows: Automate troubleshooting, refunds, and structured processes.
Integrate with External Systems: Enable Fin to pull in data or take action.
Maintain a Conversational Experience: Ensure automation remains smooth and intuitive.
Optimize Fin’s Performance with Content Suggestions
As Fin interacts with customers, it identifies knowledge gaps:
Identify Content Gaps: Update support articles accordingly.
Create New Support Content: Get AI-driven recommendations for missing topics.
Improve Resolution Rates: Enhance the knowledge base to reduce escalations.
Step 2: Test
Test Fin’s Answers
Before fully deploying Fin, it's crucial to validate Fin’s responses and ensure they align with your support expectations:
Ask Real Customer Questions: Evaluate response accuracy.
Review Fin’s Answers: Identify areas for improvement.
Refine Fin’s Configuration: Adjust guidance, content settings, and escalation rules.
✅ Compile a control set of 50 common customer queries for assessment checking whether or not Fin provided the best response.
How to Optimize Fin’s Answers
Once the test has completed, if Fin’s responses are inaccurate, here are the most common causes and how to rectify them:
Gaps in Content → Update your articles or create snippets to fill in the blanks
Lack of Guidance → Add or review the guidance you have set, ensuring it follows best practice.
Step 3: Deploy
Deploy Fin on Zendesk
Start with a controlled rollout to a subset of customers:
Select a Customer Segment: e.g., free users.
Focus on a Specific Topic: e.g., a product area.
Deployment Options
Option 1: The Intercom Messenger
✅ Fin responds to live chats in Intercom Messenger.
✅ Escalates to Zendesk agent or creates a Zendesk ticket.
Option 2: Zendesk Messaging
✅ Fin responds to live chats in Zendesk Messaging.
✅ Escalates unresolved queries to human agents.
👉 Set Up Zendesk Messaging
Option 3: Zendesk Tickets
✅ Fin replies to customer tickets in Zendesk.
✅ Escalates issues when necessary.
👉 Set Up Zendesk Tickets
Step 4: Analyze
Analyze Fin’s Performance
Tracking performance helps refine Fin’s responses and maximize efficiency:
Monitor Conversation Success Rates: Assess how well Fin handles inquiries.
Identify Areas for Improvement: Review Fin’s accuracy and refine training.
Optimize Workflows: Adjust automation and escalation rules based on insights.
For broader performance trends, use Zendesk reporting. To analyze Fin-specific performance, leverage the Fin Performance Report and Content Gaps Insights.
By following these steps, you’ll ensure Fin is effectively trained, tested, and deployed in Zendesk for seamless AI-powered support. Like a new team member, Fin will continue to learn, adapt, and improve—helping customers faster and more efficiently.