Summary
This short documentary-style video from OpenAI profiles Fyxer, an AI-powered executive assistant startup that went from zero to $30 million in annual recurring revenue in its first year by building on top of OpenAI’s model stack.
Fyxer’s founders describe the core product: an AI layer that reads, prioritizes, and drafts responses to emails and messages, so that busy professionals — the video’s flagship user is Mike, a commercial real estate manager juggling hundreds of properties and clients — can arrive at their desk with all triage already done. The team explains their technical stack in some detail: OpenAI models handle email digestion, context retrieval, re-ranking, and draft generation, with the company leaning on 500,000 hours of real executive assistant workflows to train the system’s judgment. Close collaboration with OpenAI’s engineering team on model improvements, cost reduction, and efficiency is cited as a key competitive advantage.
The video also covers the human side of building the product — from a two-person agency background to a full engineering team — and touches on why trust is the central design challenge for an AI tool that handles sensitive professional communications. For founders and product teams exploring AI agent applications in the productivity and communications space, Fyxer’s trajectory offers a concrete reference point for what GPT-based workflow automation can achieve at commercial scale.
📺 Source: OpenAI · Published September 14, 2026
🏷️ Format: Workflow Case Study
