Editor’s note: This is part of a series in which we interview active investors in artificial intelligence. Previous interviews were with investors at General Catalyst, Bessemer Venture Partners, Accel and Insight Partners.

Years before ChatGPT brought mainstream attention to AI, Index Ventures was placing bets on artificial intelligence companies ranging from San Francisco-based self-driving technology company Aurora to New York-based machine-learning platform Arthur AI.

The multistage venture firm is co-headquartered in Geneva, London and San Francisco, where it added an HQ in 2011. That move was prescient: Index now benefits from having a wide locale in which it invests. The firm raised $3.2 billion across three funds in 2021 and into 2022.

We talked with Erin Price-Wright, a partner at Index Ventures who joined the firm in 2019. Her investment focus is early-stage enterprise and AI.

AI will become like a database — where it is not a differentiator, Price-Wright predicts. “Very soon that will be true for AI.”

Price-Wright was studying for a masters at the University of Oxford in math and computer science when DeepMind, a seminal AI startup from the U.K., was acquired by Google in 2014 at a time when deep learning and neural networks were taking off. She decided against a research path and dropped out of her Ph.D. program to join the data team at data analytics company Palantir Technologies, where the big challenge was how to operationalize data for these very large companies.

She sees a lot of parallels between her work in the mid-2010s at Palantir and what enterprises are going to need over the next several years as AI is expected to become ubiquitous.

“There’s this similar challenge over the next decade of how existing enterprises with large data sets and large customer bases and complex operations start to incorporate AI to make their products better, to make their operations more efficient, to make them more streamlined, to make them faster and more nimble at decision-making,” she said.

“How effectively organizations are able to really leverage data and decision-making; I think we’re still in that transition,” she said.

In the recent before times

Before the launch of ChatGPT, artificial intelligence companies were having to beat down doors to explain what AI is, why it is powerful and why it is different from the data science wave of the past six years, she said.

Index led the $40 million Series A round in Cohere, an OpenAI competitor in September 2021 — a year before the launch of ChatGPT. Its co-founder Aidan Gomez co-authored the transformer model paper on “Attention Is All You Need” while at Google Brain.

“What ChatGPT did is shortcut that,” said Price-Wright.

The technology was not a surprise — as it had been around for a few years — but she was surprised at how quickly ChatGPT took off, and then made its way into the boardroom.

“Suddenly everybody saw, everyone could feel it really viscerally as a consumer product,” she said.

On investing in Weaviate

Price-Wright led the most recent funding in Weaviate, an open-source vector database company. It raised a $50 million Series B funding at a value of $200 million in April.

On sourcing that investment, she said Index had been spending a lot of time — close to two years — looking at vector databases.

“We saw that AI was going to be a big enabler of a new type of search for the enterprise — a semantic search which allows you to encode data into vectors and use AI to find similar information. And that’s much more powerful for many use cases than keyword search,” said Price-Wright.

The firm was also an early investor in Elastic, which enables keyword search, a predecessor to vector-powered search and vector databases.

“The event that really pushed us over the edge was ChatGPT, honestly, and it was really understanding the value of a vector database in more generative AI workflows and the importance of really understanding the embedding algorithm,” she said.

This led to two outcomes for the firm in its thinking.

“It made us realize that the market was actually way bigger than the market for enterprise search alone, if you’re thinking about all of these different types of generative AI use cases,” she said.

In particular, she added, there was a realization of “the value of having a dedicated database that could get really close to the embedding algorithms as a first-class citizen within its product, versus a bolt-on for an existing database.”

Previously, the firm was thinking about whether vector databases were going to be served by an existing database product, with MongoDB or PostgreSQL adding vector support.

Weaviate, led by co-founder and CEO Bob van Luijt, has an advantage as an open-source product to build and grow a developer community in a very efficient way and take advantage of what developers know best.

“They’re seeing a lot of adoption, especially in their hybrid cloud product, where they’re working with mature organizations rather than the AI hacker set,” Price-Wright said.

Vertical integrations

Other Index portfolio companies using AI are DeepScribe, an AI-powered product for doctors. It’s an automated medical scribe that listens to a conversation and then is able to transcribe it and convert it into an electronic health record. Index led its Series A in January 2022.

Another Index investment was in Hebbia, an AI-powered search and discovery product for the financial services industry. For a consulting project or an M&A transaction with thousands of documents to sort through and understand, Hebbia uses AI to process the information quickly and accurately.

The firm also invested in Notion, a company with strong workflows and UX that worked to incorporate AI seamlessly and naturally to the way people were using the product.

“When I talk about AI being exciting — it’s a new enabling technology. It’s something that lives within and helps provide a heartbeat for every single piece of software,” said Price-Wright.