Today we're launching QueryStory. We've been building an agentic data platform in stealth, and I finally get to show you what we've made.
The company post tells the full story. This is the personal version.
One belief
QueryStory is built on a belief I absorbed over eighteen years at Google, held so deeply there that nobody bothered to write it down:
Having data isn't the advantage. Acting on it is.
I watched that belief win over and over: tracing Operation Aurora end to end because we had the logs, all of them, queryable; founding the Threat Analysis Group; starting Chronicle from two people in a corner of the Moonshot Factory. Chronicle was that belief applied to security, treating defense as a data problem before that was obvious, and it became Google Security Operations. Astro Teller had my Chronicle co-founders and me on the Moonshot Podcast this spring to tell that story.
Stepping back
I'm a data geek. Every time a new capability arrives, my first instinct is to point it at whatever data I'm curious about. When I pointed AI at everything outside security, the tools I kept reaching for didn't exist. Cybersecurity had quietly built the most advanced data-analysis practice of any industry, because for us, trust in the data was existential. Almost none of that craft had escaped to the rest of the world, and there was a reason: everything we built assumed a curated schema that a dedicated team spent years maintaining. Security paid that tax because it had to. Most teams can't.
LLMs are the unlock. When a model can generate the query, understand messy sources, and show its work, the schema stops being the price of admission. But ideas don't apply themselves: the scaffolding and tooling security spent decades refining still have to be built for everyone else. That's what we're building.
I took it to the two people I most wanted to build with. Stan was one of Chronicle's first engineers and the reason Backstory shipped. Dave spent years inside one of the world's largest consulting firms, responsible for how more than a hundred thousand people turned data into answers. Three different roads, same conclusion: the people closest to a decision should be able to reach the data behind it.
Day one
A few days before launch, I sat down with TechCrunch's Tim Fernholz. He brought a dataset he knew the hard way: satellite launch data he had once spent over a week trying to visualize with a software engineer at his side. This time I showed him how, with just a few basic prompts, QueryStory filled the sky with every satellite launched since 2000, sources cited under the chart. That moment is the company in miniature.
Starting again, with the same belief aimed at everything else, is the most alive I've felt in my career. Every company runs on data. Almost none of them run on understanding. We're closing that gap, and as of today we're doing it in the open.
When I left Google, I called it a turning point, not an ending. Now the turn has a name.
Come see what we built: querystory.ai. And if you bring us your hardest question, we'll show you what your data has been trying to tell you.