Operator is the first item in the main navigation — select it to start new conversations and manage ongoing threads.
You can also access Fin Operator from other sections of your workspace — including Train, Test, Deploy, and Analyze. Fin Operator picks up the context of what you're looking at, so you can ask questions or request changes without switching to the Operator tab.
Fin Operator works through a simple loop: describe what you want, review what it proposes, approve and ship:
Describe your goal: Enter a natural language prompt (e.g., "Why did our resolution rate drop last week?") or upload a document like a product brief or release notes.
Let Fin Operator work: It automatically selects the tools it needs: querying data, reading conversation transcripts, browsing your knowledge base, or any combination.
Review the proposal: All changes are presented as reviewable diffs, similar to a pull request. Nothing goes live without your approval.
Approve it: Once you approve, the change goes live. Reject or edit the proposal and Fin Operator adjusts.
Schedule it (optional) Set a job to run on a recurring basis — for example, "Every Monday, analyze Fin's performance and summarize the key trends."
Worked examples
Fin Operator takes on the operational roles that keep your support setup running. Think of it less as a chatbot and more as a capable teammate who handles the management work — from analyzing what's happening in your support operation to fixing the root cause of a problem, all in a single conversation.
Here are the roles Fin Operator plays, and what each one looks like in practice.
Use case 1:
What it does: Finds content that needs updating and makes the changes — across articles, snippets, and internal articles.
Tell Fin Operator about a product or policy change — for example, "We've updated our pricing to $15/month."
Fin Operator searches your entire knowledge base semantically for any content that references pricing.
It identifies 4 articles and 2 snippets that are now outdated — including some a keyword search would have missed.
Fin Operator proposes updates to each piece of content with the correct information.
You review the diffs and approve them.
Capabilities used: Knowledge base content
Tip: A keyword search might miss content that discusses pricing indirectly. Fin Operator's semantic search catches everything — and handles the updates in bulk.
Use case 2:
What it does: Reads full conversation transcripts including Fin's reasoning, identifies what went wrong, and proposes a fix.
Share a conversation where Fin gave incorrect refund instructions.
Fin Operator reads the full transcript including Fin's internal reasoning and content sources.
It identifies the root cause: an outdated article is being used as a source, and a procedure step has an incorrect condition.
Fin Operator proposes an update to the article and a fix to the procedure step.
You review and approve both changes.
Capabilities used: Conversation debugging, knowledge base content, procedures
Tip: Without Fin Operator, you'd need to manually read the conversation, figure out which content sources Fin used (not visible in the UI), identify the issue, then navigate to the content editor and procedure builder separately. Fin Operator does this end-to-end.
Use case 3:
What it does: Analyzes conversation data to find patterns, identifies what's driving escalations, and proposes fixes — from new procedures to updated content.
Fin Operator queries conversation data to analyze conversations on the topic — volume, resolution rate, escalation patterns.
It identifies that most requests follow a predictable pattern but are being escalated because there's no procedure to handle them.
Fin Operator drafts a procedure that collects the relevant information and routes appropriately.
You review the procedure, adjust any thresholds, and approve.
Capabilities used: Reporting and analytics, conversation debugging, procedures
Tip: Fin Operator connects the data (what's happening) to the diagnosis (why it's happening) to the fix (a new procedure) — all in one conversation.
Use case 4:
What it does: Runs a broad assessment of where Fin is struggling, identifies the highest-impact gaps, and proposes content and guidance improvements.
Fin Operator analyzes your last 250 conversations, segmenting by outcome (resolved, escalated, abandoned).
It identifies that "account access" conversations have a 40% escalation rate — significantly above average.
It reads a sample of escalated account access conversations and finds Fin is missing key content about SSO setup.
Fin Operator proposes a new article covering SSO setup and a guidance rule for how Fin should handle account access questions.
You review and approve the changes.
Capabilities used: Reporting and analytics, conversation debugging, knowledge base content, guidance
Tip: This is the full data-to-action loop. Fin Operator finds the problem, diagnoses the cause, and proposes the fix — spanning four skills in a single conversation.
Start with what you need
These aren't fixed workflows, they're examples of what's possible. Fin Operator handles whatever you describe in natural language. Start with what you want to know or do, and it figures out which capabilities to use.
Need more help? Get support from our Community Forum
Find answers and get help from Intercom Support and Community Experts