Sana’s AI-Fueled Journey to a $1.1bn Exit

Long before ChatGPT made generative AI a household name, Joel Hellermark was busy figuring out how machine learning could improve human learning. In 2016, he founded Sana with a simple thesis: “By changing the way the world learns, we change the world”.

EQT Ventures invested in early 2021, one of many deals EQT has struck in the application layer of AI. By enabling employees to create, share, and apply knowledge, Sana turned individual expertise into collective intelligence – solving a major challenge for organizations seeking efficient skilling, upskilling and reskilling.

Joel and the team always saw that first product as just the beginning. By design, its foundation – indexing knowledge from every app, enabling new knowledge creation, and giving employees access to everything their company knows – was a launchpad for AI agents. And just like humans, those agents need to be onboarded to be truly effective and to work seamlessly alongside their teams.

Right now, whenever a company adds a new AI tool or agent to its stack, it has to integrate each one separately into each system of record. That introduces a lot of work and a lot of risk. But with Sana acting as a single knowledge layer, structuring all of an organization’s information just once, it enables any number of agents to be built on top of it. With Sana Agents, anyone can create expert AI agents grounded in their company’s knowledge in minutes.

Sana’s company culture

Sana’s corporate culture is as unique as its products. Each cohort of new hires at Sana spends its first week on an island outside Stockholm with one house on it. They learn about the company and working together, not by reading manuals but by solving real problems, whether it’s fixing whatever breaks in the house or answering questions about the company in whatever way they see fit. That culture of learning and iteration has become their flywheel, enabling Sana to adapt quickly as AI capabilities evolve.

That flywheel is evident in how Sana approaches its customers: meeting them where they are, then bringing them on the journey – from retrieval to orchestration, from passive knowledge to active agents. The company first sold to human resources teams looking to improve training, before expanding to democratize learning across the organization. Now it’s doing the same thing with agents, helping CIOs and leadership teams rethink how knowledge flows and is used across the enterprise by both humans and agents.

Sana has become a case study for how we evaluate early-stage AI-native founders. We look for a mix of AI native-ness (close to cutting-edge research), product design DNA (sharp understanding of user needs and the ability to turn them into intuitive experiences) and distribution (the ability to get their product into customers’ hands quickly). Joel and the team embody all three.

Access to distribution

Today, distribution matters more than ever. AI-native founders typically need less early-stage capital because they’re achieving revenue faster with smaller teams. What they really want is access to distribution at scale. That’s where investors like EQT can help. We can use the full EQT platform to give AI-native companies access to pilot customers and start building out that distribution from day one.

So far, Sana has changed how more than one million people learn and work across hundreds of leading enterprises. Workday’s recent agreement to acquire Sana means it now has access to more than 75 million users at over 65 percent of the Fortune 500. The opportunity to bring agentic knowledge to scale is enormous, reimagining how work gets done for everyone.