Doubling Down on Sana: Solving Knowledge for Enterprises

We are delighted to announce that we have led the latest financing for Sana, the global leader in applying AI to enterprise knowledge. This marks our most recent milestone with Sana, having led the company’s Series B in 2023.

Solving the Knowledge Meta Problem

From the moment we first met Sana’s founder and CEO Joel Hellermark a few years back, he has been incredibly clear on the company’s mission: to revolutionize how humans access knowledge using AI. Joel described how at a very young age he became interested in Buckminster Fuller’s idea of meta-problems—the notion that there are certain problems that, once solved, allow you to “solve everything else.”  In Joel’s mind—as in Fuller’s—the greatest meta-problem of them all is knowledge.

From what the printing press did for writing to what Google did for the internet, the tools that have accelerated access to knowledge have defined human progress. Despite humanity's unparalleled ability to create and compound knowledge, many of today's biggest unsolved problems stem from insufficient knowledge and inefficient knowledge distribution.

When most of the existing digital knowledge systems were invented, businesses were just moving to the cloud. Company knowledge got scattered across dozens of apps, so enterprise search tools emerged to help us find files—but the results kept leaving us disappointed.

Fast forward to today, and businesses are moving at lightning speed, yet employees still spend hours searching for information, waiting for answers from colleagues, and completing mundane, manual tasks.

Knowledge workers don’t just want to find knowledge quickly—they want to apply it quickly. Generative AI makes that possible. Traditional knowledge, learning, and enterprise search systems were built for a time that’s passed.

Sana has always known this. The company first emerged as an AI-powered learning platform before introducing its AI assistant platform. When we first invested in 2023, we were impressed by Sana’s ability to challenge the status quo of existing knowledge management and enterprise search categories. Years before GPT-3.5 and the birth of ChatGPT, Sana had already been utilizing OpenAI’s models to help organizations generate new and repurpose existing knowledge more effectively, and we believe was the first to enable semantic search and natural language responses within a learning platform. Companies took notice. Their customers include enterprises across a wide-range of industries—from Merck and Electrolux to Hinge Health and Svea Solar.

Ultimately, AI and large language models (LLMs) are only as good as the context they are given. It’s our belief that—without solving the knowledge creation and dissemination problems that enterprises already have—broad-based enterprise AI adoption will always face significant barriers. That’s why we are excited by Sana’s unique positioning as a company with knowledge improvement at its core.

Today, Sana’s AI products allow you to index knowledge from all of your apps, create new knowledge in a range of interactive formats, and access everything your company knows through search and chat. And they’re just getting started.

Pioneering the 3 Arcs of Enterprise AI

Alongside the recent financing, Sana is announcing a range of new features that enable the company to deliver on what it describes as the "3 arcs of AI" for the enterprise:

• AI that searches and finds information: Today, most company knowledge is scattered across apps, stuck in Slack threads and employee's brains, or lost through verbal conversations that never get captured. Sana’s enterprise search capability provides universal search across all company knowledge, data, and apps, including meetings. Its latest release includes enhanced ranking across structured and unstructured data.

• AI assistants that answer questions through chat: Sana is launching a new UI to make it even easier for companies to create, fine-tune, and interact with AI assistants grounded in company data, and all from one platform. With this release, a single AI assistant will be able to execute multiple tasks and actions across files, apps, and meetings in parallel. 

• AI agents that act and problem-solve both with and for you: Beyond assistants, Sana is announcing features for organizations to build custom AI agents for any team or workflow, without writing code. Powered by Sana’s state-of-the-art RAG and its agent solution, A-4, these agents will collaborate to solve complex multi-step problems, complete tasks in other apps, and self-critique their output until they’ve resolved the task at hand. They will also be multimodal, capable of dynamically generating the most appropriate UI required to solve the problem. 

Existing customers are seeing the business value of Sana across multiple use cases, including : 95% time savings from finding complex product information and up to 90% productivity gains from automating complex workflows for industrial companies; more than 60% in time savings from automating legal admin and 200% increased efficiency for legal firms; and more than 10 hours per week time savings in new hire onboarding for internal HR teams. [1]

We’ve also been impressed by the early proof points from Sana’s product-led motion. In April this year, the company released a Free tier to its AI assistant platform and has seen 100,000 [2] workspaces created in just six months.

Looking Ahead: Building the UI for AI for the Next Billion Users

Sana believes that the barrier to broader enterprise AI adoption isn't a technology problem—it's a UI problem, and we agree.

Here, we believe Sana's Scandinavian design ethos—being founded and headquartered in Sweden—is a unique advantage and sets it apart from incumbent competitors like Microsoft, especially when viewed in conjunction with its AI R&D. Its new hires like Erik Olmers a designer from Apple and Oscar Täckström, Chief Scientist and former a researcher at Google AI, are a signal of the investment Sana intends to make in this area with the new funding.

Most users struggle to prompt LLMs, and chat is often not the best interface for complex tasks. The next AI interface will morph into what we need it to be, anticipate what we need to surface suggestions proactively, and be ubiquitous.

Sana also believes that AI needs to be easier to use and easier to deploy. It’s committed to making the ability to deploy custom AI assistants and agents fast and simple, without compromising on enterprise-grade security, privacy and real-time permissions.

By solving both problems, Sana unlocks the potential for companies to scale from a few enterprise use cases to thousands, without the costs of building and maintaining custom-built, in-house solutions.

—

In just a little more than a year since we first invested, Sana has rapidly scaled both its product offering and its global footprint, launching new offices in London and New York, and expanding its customer base to 100s of customers across multiple geographies across the globe. Sana is always looking to add more brilliant people to its team—you can see their open positions here.

When we initially partnered with Sana, Joel’s rare combination of technical and product acumen paired with a strong commercial sense was a determining factor in our decision. Since then, we have been impressed by Sana’s unwavering commitment to talent density as it scales. Every new employee is handpicked and goes through a rigorous hiring and onboarding process. Product velocity, already high at the time of our investment, also accelerated and we believe it’s still day one.

At NEA, we are thrilled to continue supporting Sana on this journey. The progress the team has made in just a year has been extraordinary, and we are more confident than ever in their ability to build the ultimate AI user interface for enterprises. We can’t wait to see what’s next!