Decagon's Playbook: Why Enterprise AI Agents Are Shifting to 90% Open-Source
a16z · youtube · 2026-07-31
a16z Podcast hosts the co-founders of Decagon, a rapidly growing enterprise AI agent company, to discuss the realities of building production-grade AI at scale.
Key Insights
- Embracing Open-Source & Fine-Tuning: Decagon transitioned 90% of its inference workload from frontier APIs to open-source models. They argue that fine-tuned smaller models often outperform larger models in specific tasks while optimizing latency and cost.
- Beyond API Plug-and-Play: Production-grade enterprise AI requires far more than just calling frontier models, relying heavily on robust evaluation systems, fine-tuning, and forward-deployed engineering.
- Agents Building Agents: The conversation covers their internal tool, Duet Autopilot, which automates the creation and optimization of agent workflows.
- Business & Competition: They analyze "glass box vs. black box" approaches in enterprise sales and discuss why application-layer startups will continue to thrive alongside foundation model labs.
Related event: Decagon Shares Enterprise AI Customer Service Agent Insights(2 posts)→
More from coding & agent
- Comparing 4 AI Agent Observability Platforms: Logfire Takes the Crown with Native SQL — samuelcolvin · 2026-07-31
- Testing Opus 5: Building a ThreeJS App Zooming Across Nine Powers of Ten — omarsar0 · 2026-07-31
- Developer Shares Codex Micro Workflow: Monitor AI Agents Between Workout Sets — rudrank · 2026-07-31
- Open-Source Autonomous OS Deployed on Reachy Mini Robot Hardware — dee_hw · 2026-07-31
- Rails Solid Queue 1.6 Released: Optimized for Long-Lived LLM Jobs — kieranklaassen · 2026-07-31
- Inference APIs Returning Encrypted Blobs Break Session Portability — petrusenko_max · 2026-07-31