While Sifted readers might be most familiar with Thomson Reuters’ news service, journalism today only accounts for “around 10%” of the company’s business, explains chief product officer David Wong.
It’s primarily a provider of data, information and advisory services to clients in the legal, tax and accounting sectors — and it’s now planning on using AI to double down on these offerings. And it’s putting serious capital behind the plan: it has around $10bn to invest in and acquire startups which can speed up Thomson Reuters’ product innovation.
“In the past couple of years Thomson Reuters has been in the middle of a pretty important transition and transformation of our business. AI is providing a huge opportunity for us to become much more of a technology and a software company, instead of a content and data company,” says Wong. “I think that’s a huge opportunity for how we work with startups.”
So far this year in Europe, Thomson Reuters has acquired Swedish invoicing and tax software startup Pagero and UK-based legal AI startup Safe Sign Technologies, as well as US-based tax and accounting AI startup Materia.
So how is the company planning on spending the rest of that $10bn, and how startups can best work with it?
Domain expertise
Wong explains that Thomson Reuters uses a “build, partner and buy” strategy for product development, and looks to acquire startups “where we think that there could be a new capability, or that we can move faster [by acquiring].”
Cambridge-founded Safe Sign Technologies was a good example of this, he says, as the startup had cracked technical problems around building large language model (LLM) technology for the legal profession.
“They had focused their efforts on discovering new ways to create these specialised, domain-specific language models, and they had cracked a few problems that we had not cracked yet,” Wong tells Sifted, adding that Thomson Reuters was able to solve a problem for the startup, by giving its research efforts access to more data and readily available legal expertise.
He says that Thomson Reuters will continue to eye acquisition and investment opportunities in startups that are working on applying GenAI technology to very specialised, domain-specific problems.
“When concerted effort is put against trying to solve a particular narrow problem, you know — let’s say, trying to evaluate a particular class of contracts, or to draft a particular type of agreement — that actually can rival and exceed human quality. That’s one theme: which is specialising and building domain specific capabilities,” Wong tells Sifted.
Agents
Thomson Reuters is largely looking to partner with or buy AI startups that help solve its clients’ problems around knowledge retrieval and content creation, but is also interested in investing in startups that are working on “agentic” AI capabilities. What this means is technology that can automate whole tasks, rather than having to be prompted every step of the way.
“More and more tasks are going to be augmented or automated on behalf of the lawyer, the accountant, the tax preparer,” says Wong. “It’s about being the orchestrator of work, and being able to then effectively use the AI to be your team. That’s the trend that we’re seeing right now.”
“That’s one of the reasons why we invested in Materia, because they had cracked certain types of agentic flows… We’re very, very interested in taking a look at those types of investments.”
How to work with Thomson Reuters
Wong tells Sifted that the best way for startups to begin a partnership with Thomson Reuters is to reach out to the team at Thomson Reuters Ventures — the company’s early-stage investment fund.
“That’s probably the widest part of our funnel, just looking at how we set up relationships with startups,” he says.
Wong recommends that the next port of call for startups is to reach out to Thomson Reuters’ product teams, to begin demonstrating that there is a strong fit ahead of an acquisition.
“The ideal situations for us are where we partner first and establish how we work together with a startup to see if there is a complementary solution,” he says. “It’s always best to be able to put something in the hands of customers where they can see the power of the startup’s and Thomson Reuters’ technology together.”
He adds that one advantage he sees in European startups, compared to US counterparts, is that they tend to think more globally, rather than solely focusing on their home market.
“US-based startups are generally satisfied with the US market as a way of defining their TAM [total addressable market] and so, as a result, are US-centric. European and other non-US startups think about the whole world, and that’s really valuable for Thomson Reuters, because we’re a global company.”
Read the orginal article: https://sifted.eu/articles/thomson-reuters-ai-strategy/
