← All speakers

Bio, Work & Ideas

Ryan Cooke

Conference affiliation: WorkOS · 2026

Ryan Cooke presented WorkOS’s approach to AI-assisted product engineering at AI Engineer World’s Fair 2026, where he was identified as an engineer at WorkOS. His talk, “No, That’s Not a Software Factory,” argues that engineering automation should be judged by customer delivery and reliability rather than the volume of generated code or pull requests.

Cooke describes how WorkOS embeds planning, project tracking and support triage into its software factory. TARS integrates with Slack, Linear and GitHub, using webhooks to track progress, pick up newly unblocked tickets and reassess plans for missing work. A PM agent drafts Hilltop product documents and breaks reviewed work into tickets, giving engineers a starting point they can refine. Horizon provides infrastructure orchestration, while an internal MCP gateway connects tools and supplies guidance on their use, including access to Snowflake data for product and customer analysis. Cooke also discusses automated bug and support triage, developing company-owned sandbox infrastructure, and plans for shared organizational memory across AI tools. He emphasizes human review, defect rates, recovery time and voluntary engineer adoption as signals of value.

1 conference talk

Key ideas

Scroll to read ↓

WorkOS’s first sandbox could produce pull requests, but Ryan Cooke found little difference from engineers using coding agents locally. TARS and Horizon extend automation into planning, review, project coordination and context gathering—the work that turns code into a product.

  • A sandbox can generate code without improving product delivery. WorkOS’s initial experience pushed it toward automating planning and coordination alongside implementation.
    3:21 ↗
  • Ticket dependencies and completion webhooks let TARS continue a project without a fresh prompt at every handoff; requested reevaluation helps expose missing work.
    5:30 ↗
  • Draft Hilltop documents reduce setup work and give engineers something to refine, but scope control remains an important human contribution.
    8:14 ↗
  • An internal MCP gateway becomes more useful when it explains where information lives and when to query it. WorkOS reuses that guidance for coding, data analysis and customer analysis.
    12:15 ↗
  • Evaluate customer delivery alongside defect rate, recovery time and voluntary adoption. Company-wide memory and session-driven improvement are future directions; authorization remains unresolved.
    15:27 ↗

References