Product Ops With AI Agents
6 posts
Running a SaaS product with an AI agent: feedback, support, tasks and shipping, built on checked vendor documentation and real runs.

A real run from a customer request to the "It's live" email in LotsTeam: post, task, link, changelog and what the product emailed. Steps and limits.
Who the agent acts as, what it can write or delete, and which plans include it, for Canny, Featurebase, Help Scout and more. Checked 3 November 2026.
A copy-paste feature request form template with every field explained, three filled-in examples, and how to collect, vote on and deduplicate requests.
How to prioritize feature requests with an AI agent: a real run scoring tasks on reach, pain, revenue and effort, with the prompt and tool output.
I had an AI agent triage 14 feedback posts through the LotsTeam MCP: duplicates grouped, 7 tasks created, replies posted. The prompt and real output.
Which support desks and issue trackers have an MCP server: Intercom, Help Scout, Zendesk, Linear, Jira, GitHub, Asana and ClickUp, checked in their own docs.
No more posts to load