How Fyxer built an AI executive assistant people trust

3 min video3 key momentsWatch original

⚡

TL;DR

Fyxer built an AI executive assistant that saves busy professionals hours daily by automatically prioritizing emails and conversations — they hit $30M ARR in year one using OpenAI's models.

Key Insights

Lost 1-2 hours daily — A commercial real estate manager handling hundreds of properties was losing 1-2 hours daily to email catch-up after work, killing family time — Fyxer's triage cut that to minutes.

500,000 hours in months — OpenAI's models let Fyxer learn from 500,000 hours of executive assistant workflows in months instead of years, accelerating their path to $30M ARR.

ChatGPT is the second-most-used AI tool by Fyxer customers after their own product — showing they're stacking OpenAI's capabilities rather than replacing them.

Want this for every new video OpenAI posts? Brevyd summarizes each upload automatically, the morning it drops.

Deep Dive

The problem Fyxer solves

Mike, a commercial real estate manager juggling hundreds of properties and people, spends 1-2 hours every evening plowing through email instead of being present with his family. He can't respond while driving. Fyxer's core mission is to hand him a pre-sorted inbox when he opens his laptop — all emails read, ranked by urgency, with only what actually needs him at the top. The team wanted to move Mike from reactive catch-up mode to proactive decision-making. One real customer told them the product saved him from divorce.

Building on OpenAI's models

Fyxer's engineering team — Joey, Eleonora, and David — work closely with OpenAI. On his first day, the CTO told Eleonora her job was to make the auto-generated email drafts better. They use OpenAI models across the entire pipeline: digesting emails to understand their intent, pulling in relevant context, and re-ranking which messages deserve top billing. The team credits OpenAI not just for the raw model quality but for hands-on collaboration on costs, efficiency, and technical direction. The breakthrough came when GPT-3 arrived and they got exactly the output they expected.

Scaling from zero to $30M

Before OpenAI's models, Fyxer was manually building an executive assistant agency — slow, unscalable. GPT-3 changed everything. They could suddenly learn from 500,000 hours of real EA workflows in a matter of months instead of years. That compressed timeline let them go from zero to $30M in annual recurring revenue in their first year. The founders emphasize that the hard work was still understanding their customers deeply — OpenAI's platform just made it possible to translate that learning directly into how the models actually behave. ChatGPT itself is the second-most-used AI tool among their users, suggesting Fyxer is complementing rather than replacing existing AI consumption.

Takeaways

✓If you're drowning in email admin, delegating prioritization to an AI saves 1-2 hours daily — enough to reclaim family time or deep work.
✓Building a successful AI product means learning your customer obsessively first, then using frontier models to scale that learning — not the other way around.
✓ChatGPT or GPT-4 on top of a specialized AI product is additive, not redundant — stack them rather than assume one replaces the other.

Key moments

0:02Divorce-saving claim

“This year, I think you've saved me from getting a divorce.”

1:30GPT-3 breakthrough

“With the release of GPT-3, it was the moment where I just like got exactly the answer back that I expected.”

2:26$30M in first year

“OpenAI actually made Fixer possible. It allowed us to learn from the 500,000 hours of EA workflows in just a matter of months. And the result of that was in our first year we went from zero to 30 million in ARR.”

You just read one. Brevyd does this for every upload.

Follow OpenAI and every new video comes back as a summary like this, in your morning briefing. No watching required.