Supra Insider
Supra Insider
#71: How PMs can get leverage via agents and MCPs in Cursor | Amir M (Cofounder @ Humblytics)
2
0:00
-1:03:18

#71: How PMs can get leverage via agents and MCPs in Cursor | Amir M (Cofounder @ Humblytics)

A deep dive into using Cursor and MCPs for PRDs, automated QA, and scaling product workflows without scaling the team
2

Listen now: Spotify, Apple and YouTube

What if you could cut your QA cycles from days to minutes—and draft PRDs that actually update themselves as your product evolves?

In this episode of Supra Insider, Marc and Ben sit down with Amir M, cofounder of Humblytics, to explore how he’s running a two-person startup across engineering, QA, and product using Cursor and Model Context Protocols (MCPs). Amir shares how he builds context-rich workflows, turns documentation into living systems, and uses agentic tools like Firecrawl and Playwright to automate the “boring” but critical parts of product development.

If you’ve been curious about how to bring AI deeper into your product org—not just for brainstorming but for end-to-end execution—this conversation is packed with practical demos and mindsets you can apply today.

All episodes of the podcast are also available on Spotify, Apple and YouTube.


New to the pod? Subscribe below to get the next episode in your inbox 👇


In this episode, we covered the following topics:

  • (01:11) Amir explains the tiers of AI fluency for PMs and how constraints of a small team pushed him to adopt Cursor beyond coding.

  • (03:50) Why Anthropic’s rebranding of MCPs as “tools” or “connectors” makes them more approachable for non-technical teams.

  • (06:51) Amir’s core MCP stack: Firecrawl for scraping, Playwright for agentic browser actions, and Perplexity for research.

  • (12:31) The future of browsers in an MCP world and the rise of Generative Engine Optimization (GEO) as the new SEO.

  • (21:33) Filtering noise vs. signal in AI trends and why true experts in this space are “tinkerers” rather than capital-E Experts.

  • (30:01) Live demo: Using Cursor to auto-generate a PRD for a new “no-code testing” feature, leveraging context from documentation.

  • (45:15) How custom modes, rules, and context engineering combine to create high-quality PRDs and QA plans.

  • (51:15) Automated QA in action: Using Playwright MCP to run test plans agentically—no manual clicking required.

  • (57:14) Running background QA agents and tying results into production workflows with Slack alerts.

  • And more!


Links:


If you’re a new subscriber, we encourage you to check out some of the recent episodes you might have missed:

And here are some early favorites:


To support the podcast, please check out the links below:


Discussion about this episode

User's avatar