Beyond the model: How we’re shaping Mixpanel Agent around the way people work

Mixpanel reports that use of its AI Agent increased tenfold after launch, driven by customer feedback, product analytics, session replay analysis, and retention-focused measures rather than engagement alone. Although fewer than 1% of agent interactions receive direct feedback, users who responded reported an 80% weekly happiness score following the agent’s June 2026 general availability, while constructive feedback helped identify priorities. Improvements included an Artifacts section that exposes all reports and components generated in a response, clickable report titles that increased clickthrough rates by 10%, documentation-grounded answers to reduce inaccurate product information, and context inheritance so reports created within Boards retain the Board’s date range or edit an open report. Mixpanel also shifted longer-running tasks to background execution, allowing users to leave or close tabs without losing results; 7% of turns now run asynchronously. Customer demand additionally led to support for features such as Custom Properties, Custom Buckets, and Lookup Tables, while planned work includes bulk editing, proactive insights, and expanded context and customization.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.