Thomson Reuters is a venerable news and information organization, with its historical roots stretching all the way back to the 19th century. The two companies merged in 2008 and provide a combination of news and specialized information in areas like law, trade and accounting.
The organization processes a ton of information every year, relying on a staff of 27,000 subject experts and journalists around the world to generate a variety of content. As generative AI has emerged in recent months, it would surely be tempting to use it in the newsroom, as other news organizations have done, and see this capability as an opportunity to reduce staff, cut costs and automate, automate, automate.
While the company sees the benefits of AI for both its employees and customers, it is not in the worker replacement camp, at least not yet. Instead, it sees AI as a way to help customers find information faster, and help its employees operate more efficiently, removing the mundane parts of the job so people can do what they do best.
It would be easy to think that an organization as old as Thomson Reuters would simply dismiss technology like generative AI, but the company tells TechCrunch+ that it is all in when it comes to the latest technology, as it looks for ways to improve and modernize its operations.
The people part
Chief people officer Mary Alice Vuicic, says Thomson Reuters sees automation as only part of the story, and if you concentrate on that, you may miss some of AI’s biggest benefits.
“We think AI is a phenomenal opportunity for the professionals we serve through our products, and equally internally for our colleagues,” Vuicic told TechCrunch+. “We think it’s a tool for augmenting the potential of our colleagues in new ways, helping them do work better, faster, more effectively.”
That said, she also recognizes that large language models (LLMs) do not always provide perfect answers, and Thomson Reuters is already relying on internal expertise to help correct the models.
As an example, Thomson Reuters is using in-house legal experts to test and refine the answers coming from LLMs. “We have a significant population of attorney editors who leverage their expertise to write content to inform the legal perspectives and decision-making that feeds into our products. This has been instrumental in training the AI in our legal tech, and we’re using them as customer zero users validating and correcting generated answers,” she said.
As the company is working to expose AI to the workforce, it had a learning day in April dedicated to AI, which attracted 6,000 employees to the live event, with thousands more watching later on video. “The focus was on doubling down on learning about how this is going to impact us, and how it can be leveraged for customers, as well as how it can be leveraged internally. And then more broadly, just understanding the technology,” Vuicic said. “We’ve launched AI guiding principles and talent principles, so people understand responsible use of AI and the responsibility that we have in the role that we play in the market as well as as employers.”
She said there is also a big communications component, just letting people know how the company is planning on using AI, and as part of that strategy there is an internal portal devoted to the subject where people can go and learn what is happening around AI.
Making the investment
Thomson Reuters has invested heavily in AI, committing $100 million a year toward the technology. It recently announced the $650 million acquisition of Casetext, an AI-driven legal tech startup. “The acquisition of Casetext is another step in our ‘build, partner and buy’ strategy to bring generative AI solutions to our customers,” Thomson Reuters CEO Steve Hasker said in a statement accompanying the announcement.
Thomson Reuters also teamed up with Microsoft Office Copilot, where lawyers using Thomson Reuters legal products can ask questions to access legal content using a natural language query without leaving Office to get the information.
“That’s on the legal side, but we actually see opportunities across much of our product portfolio for this. What’s critical is that our colleagues understand AI and are learning rapidly,” Vuicic said.
The company thinks that providing a way for employees to learn in a secure way will give it a competitive advantage and lets people understand the purpose of this technology to augment their jobs. “A key focus for us is for people to see the opportunity, not retrench out of fear, but see the opportunity to be at the forefront of this.”
With AI changing so rapidly, the company is keeping an open mind when it comes to which LLMs it uses, this in spite of the partnership with Microsoft (which in turn has a partnership with OpenAI). Neither are they thinking of building their own LLM at this time.
Ultimately, Vuicic sees AI as a way to enhance human workers and customers interacting with Thomson Reuters content to use the content more effectively, or to make workers more productive and efficient.
She sees a real opportunity to take the intellectual capital that was being used to complete mundane tasks and channel it toward innovation, R&D and business growth. “If companies effectively manage the adoption and accelerate the adoption of AI through a human-centered approach, they will reap the real benefits of this and have a competitive advantage,” she said.