Chats with Kent C. Dodds Foundations, feedback, and agents — Dillon Mulroy on product at Cloudflare
Apr 1, 2026
Dillon Mulroy, principal engineer at Cloudflare who builds AI agents and developer platforms. He talks about using real user feedback and observability to shape products. He explains sitting with support, scoping small when AI speeds coding, and why strong foundations and paging matter. He also covers reviewing agent-written code and building cross-functional relationships to prioritize fixes.
AI Snips
Chapters
Transcript
Episode notes
Dogfood Sister Team Products Internally
- Do dogfood and be 'customer zero' for sibling platform teams to provide early feedback before public releases.
- Dillon's agent experience team acts as customer zero for the agents SDK and AI platform at Cloudflare to influence product design.
Get Feedback By Being Where Your Users Are
- Do get direct user feedback by being where your users are and forcing yourself into user conversations.
- Dillon recommends checking forum posts, social streams, and sitting down to watch users use the product to surface real paper cuts and edge cases.
Use Support Tickets As HighImpact Product Signal
- Do build strong relationships with customer support and treat their tickets as a product feedback goldmine.
- Dillon says sit with support, triage cases, and build tools that reduce recurring escalations to reveal high-impact product fixes.