Accenture takes stake in generative AI firm Stardog

Accenture says it wants to 'nurture' the development AI business use cases

Global solution provider Accenture has revealed it has invested in AI firm Stardog.

The company has developed technology that combines data from multiple sources and makes it machine-understandable without changing the underlying data. The aim is to allow such data to be easily searched, and it is being touted as a way to improve searches using generative AI.

Accenture is not acquiring Stardog, but is instead making a strategic minority investment in the company via its Accenture Ventures, said Teresa Tung, chief technologist at Accenture Cloud First.

"We believe this is such a strategic technology that we want to get in with the startups so that we can shape their product to be best for helping us with our clients," Tung told CRN. "We have a venture arm. And we have a Project Spotlight which is doing just that, picking out some of these very strategic technologies that we need to nurture and grow."

Accenture, which currently partners with Arlington, Va-based Stardog, did not disclose the size of its investment in the company.

The investment doesn't give Accenture first rights to Stardog's technology, Tung said.

"But we do get a chance to influence their roadmap," she said. "As an investor, we get to see a bit more. And we get to help them influence and shape what they're doing, and to nurture these technologies that we think are really critical."

Unlike the common data lake-centric take on data, modern data stacks that are emerging today are all about connecting across data lakes, partners, and systems, Tung said.

"Stardog's knowledge graph gives you that connectivity map that indexes over all that data," she said. "We think that's just really pivotal. So we really want to be able to make sure that grows."

It's important to bring those different data sources together, Tung said.

"Multi-lake, multi-system, multi-cloud, multi-partner, because of data sovereignty, that's just how data has grown," she said. "I always say I wish it could all be central in the same place by the same team. But I wish for a lot of things."

Because of data sovereignty, data remains stored in different stacks, but businesses need to access it like they do on the internet, Tung said.

"Internet searches are using something similar to what Stardog provides, a domain knowledge graph," she said. "So when we search for things like ‘Accenture,' it's not just giving you the indexes and the links, it's taking it and it's formatting ‘cards.' You have Accenture, you have the map closest to you, there's a little card. There's another little card that says who are competitors. Tough questions people ask about Accenture. And that's all from the knowledge graph. It's giving you that data into context."

Stardog does not use generative AI as part of its technology, Tung said.

"I would say that it's more of an enabler for generative AI," she said. "So the large language models that underpin generative AI, I would say that's not what Stardog does. Stardog adds that domain context."

Tung gave an example of how the Stardog technology might work.

"So if I were a retail company and I had my own knowledge graph with my products and my customers, and wanted to give my customers a way to say something like, ‘I'm going on vacation, what should I buy? What should I pack?' You could use the large language model to figure out that people who go on vacation in the summer need swimsuits and shorts. And then you could map it with my own data within Stardog to say, ‘Well, our products happen to have swimsuits, and Theresa is a customer and she likes blue. So let's serve her blue swimsuits.'

Stardog has raised a total of $23.3m in funding since its founding in 2015 and has about 100 employees, according to Crunchbase and LinkedIn.