← Front page

The a16z Show · Friday, July 31, 2026

Decagon Shifts 90% of AI Workflow to Open Source for Latency and Control

Decagon, a company focused on enterprise AI deployment, has transitioned 90% of its AI workflow to open source models, primarily to address latency issues for its voice agents. Co-founders Jesse Zheng and Ashwin Srinivasa discussed this shift, highlighting that fine-tuned open-source models can outperform larger, closed-source models on specific tasks.

personJesse ZhengpersonAshwin SrinivasacompanyDecagon

The tape

3 quotes
So today, 90% of our workflow is on open source.
Ashwin Srinivasa
On the specific task we want them to do, they actually outperform the large, smart state of the art model.
Ashwin Srinivasa
The thing that we built was not an agent that does customer support well, but rather an agent that follows business process well.
Ashwin Srinivasa
Heard on The a16z Show — “How Enterprise AI Really Gets Deployed, published Friday, July 31, 2026. Heardvine summarizes and quotes with attribution and timestamps, and links to the original everywhere.
Transcribed via Gemini audio transcription · $0.07
Decagon Shifts 90% of AI Workflow to Open Source for Latency and Control — Heardvine