Managed service news
Joseph F. Kovar
“We have a chance to influence their roadmap. As an investor, we can see a little more. We can give, shape, and help nurture these technologies that we think are really important,” said Teresa Tung, Principal Technologist, Accenture Cloud First.
Global solutions provider Accenture on Wednesday announced an investment in Stardog, which develops technology that develops a knowledge graph of data that generative AI can use for detailed searches.
Stardog is the developer of Stardog Enterprise Knowledge Graphs, a technology that combines data from multiple sources and makes it understandable by machines without changing the underlying data. The aim is to make such data easily searchable, and it’s touted as a way to improve search using generative AI.
Rather than buying Stardog, Accenture is making a strategic minority investment in the company through Accenture Ventures, said Teresa Tung, chief technologist at Accenture Cloud First.
[Related: Accenture CEO Julie Sweet On Layoffs, Belt-Tightening, And Going After ‘Structural Costs’]
“We believe this is a very strategic technology, so we want to work with startups so that they can come up with the perfect product to support our clients,” Tung said. told CRN. “We have a venture department. increase.”
Accenture is currently partnering with Arlington, Virginia-based Stardog, but has not disclosed the size of its investment in the company.
Tan said the investment does not give Accenture any preference over Stardog’s technology.
“But we have an opportunity to influence their roadmap,” she said. “As investors, we can see a little more. can help you do it.”
Unlike the typical data lake-centric mindset for data, all modern data stacks emerging today are about connecting data lakes, partners, and systems, said Tung.
“Stardog’s Knowledge Graph provides a connectivity map that indexes all your data,” she said. “We think it’s really extremely important, so we really want to make sure it grows.”
Integrating these disparate data sources is key, says Tung.
“Thanks to data sovereignty, multi-lake, multi-system, multi-cloud, multi-partner, that’s exactly how data has grown,” she said. “I always wish the same team could be in the same place and centered on everything. But I wish a lot.”
For data sovereignty, data remains stored in different stacks, but businesses should access it the same way they do on the internet, Tung said.
“Internet searches use something similar to what Stardog offers: a domain knowledge graph,” she said. “So when you search for something like ‘Accenture’, it doesn’t just give you an index or a link, it takes the information and formats it into a ‘card.’ There’s one more little card that shows who the competition is: The tough questions people have about Accenture.That’s the explanation from the Knowledge Graph.You can relate the data to the context.”
Stardog does not use generative AI as part of its technology, Tung said.
“This is more of an enabler for generative AI,” she said. “Thus, I would say that large-scale language models underpinning generative AI is not what Stardog does. Stardog adds context to that domain her.”
Tung gave an example of how the Stardog technology works.
“I mean, let’s say I’m a retail company, and I have my own knowledge graph about products and customers, and I want to give my customers a way to say, ‘I’m going on vacation. What should I buy?” What should I pack?” Using the large language model, we know that a summer vacationer needs a bathing suit and shorts. can be mapped with my own data within Stardog to say, “Our product happens to have swimwear, Teresa is a customer, and she likes blue.” Let’s serve her blue swimsuit.
According to Crunchbase and LinkedIn, Stardog has raised a total of $23.3 million since its founding in 2015 and has about 100 employees.
Joseph F. Kovar
