Business teams now ask 73% of the questions in Delphina. CEOs at more than half of our customers are active users. We've been betting on this since Uber: when everyone can get trusted analytics, demand skyrockets. It's happening faster than even we expected.
We're raising our ambition 10x. Today we're announcing new funding co-led by Radical Ventures and Costanoa Ventures, bringing total funding to $15M. We're also launching the Context Factory, staffed with Delphina agents that build and maintain the business knowledge AI needs to get your numbers right.
Customer momentum
The adoption has been electric. Our top users are in Delphina every day, including weekends, for the last 3 months. Three quarters of usage comes from executives and operators outside the data team. As one leader put it, "the demand for accurate data is viral, it's like early Facebook."
Executives and ICs alike start their day with Delphina at leading companies across industries, including Coursera, Substack, LATAM Airlines, and more.
The unfair advantage of being data-driven
At Uber we were building Michelangelo and leading global pricing data science.
There, CEO and founder Travis Kalanick pushed everyone at the company, including himself, to be hands on with data. That direct line to the numbers was a real advantage as Uber grew from $1B to more than $100B in gross bookings. Yet it also took hundreds of data engineers to keep it from collapsing into chaos; very few other companies could afford this.
So most settled, including many ambitious ones. They'd get dashboards and a ticket queue, and maybe 10% of the value their data should deliver. Many of these companies ended up data obsessed, but not data-driven.
AI changes the math. Everyone in a company now can (and needs to be) a consumer of data, and agents can go deeper and wider than any analyst. But it's only useful if the answers are right.
The hard part isn't the model
The models are good enough. What they don't know is your business: how "active user" changed last quarter, what business priorities were last month, what your best analyst knows that nobody wrote down.
The standard fix is to try to write it all down by hand. Semantic layers, glossaries, prompt libraries. The industry is littered with 18-month projects that never finished, and the ones that did went stale the next time the business changed. We call it the data context trap. We don't think "context engineer" should be anyone's full-time job. That approach won't work.
Introducing the Context Factory
The Context Factory flips the work. Agents build your company's context from what you already have, in the warehouse and in Slack, Notion, Drive and GitHub, and reach a working baseline in hours. Everything is code, with secure sandboxing, version control, and governance built in.
Then they keep going through the night. When an answer is uncertain or two definitions conflict, they trace the gap to its source, draft the fix, and send it to the person who knows. Once it's approved, every future answer inherits it. The context is code, versioned and governed, and it's yours: readable files, independent of any model.
This is where data teams win. Instead of working a ticket queue, they become knowledge captains: the people who decide how the business sees its numbers, with agents doing the heavy lifting. Their judgment shows up in every answer, for everyone, every day. That's more leverage than a data team has ever had.
With Delphina we're seeing the emergence of a post-modern data stack, with AI, code and governed context at the center instead of one warehouse.
What it looks like
- Substack: More than 70% of the company uses Delphina regularly, including CEO Chris Best: "It's changed how we make decisions."
- Medely: The Q3 board deck was 75% drafted the day after quarter end, before the books closed. It used to take more than 30 days.
- LATAM Airlines: Chief Data & Engineering Officer Andres Bucchi describes Delphina as "turning every single data point into enterprise value."
- BaseCamp Franchising: The co-CEO built a report in 40 minutes that had been stuck for six months. It's now a company-level initiative.
- Gen Digital: From zero to more than 9,000 AI analytics conversations a month in six months.
What's next
Thank you to Rob at Radical, Tony at Costanoa, and our angels, including Fei-Fei Li, Nobel laureate Guido Imbens, and early employees from Anthropic and OpenAI. This round lets us scale the Context Factory, grow our go-to-market and customer engineering teams, and take Delphina company-wide at more enterprises.
Every company now runs on two workforces, people and agents. They need one shared, trusted understanding of how the business works. That's the foundation of the AI-native company, and it's what we're building.
If your company has more questions than answers, let's talk. If you want to build this with us, we're hiring.
— Jeremy and Duncan