The company also got a future chief technology officer out of the deal.

Joel Hron, who had served as CTO of ThoughtTrace at the time of the acquisition, came into the fold at Thomson Reuters as a VP of technology to spearhead integration into the larger company’s ecosystem that includes the Reuters news agency and various tools sold to legal, tax, and other corporate professionals.

By 2023, Hron became head of AI and Thomson Reuters Labs, a decades-old research and development group whose work was dramatically changing as the generative AI boom was leading companies to rethink their approach on AI. During Hron’s tenure, TR Labs accelerated hiring and launched seven generative AI products in 18 months, including AI assistants for legal research and contract drafting.

That group still reports to Hron, who ascended to the CTO role in July 2024. Using the lessons he learned as an outsider who joined Thomson Reuters via an acquired startup, Hron is now helping to steer an aggressive M&A strategy that will continue to see more newly acquired technologies and teams added to the roster.

Since 2019, Thomson Reuters has spent $4.2 billion on acquisitions, adding e-invoicing provider Pagero, and automation software providers SafeSend and SurePrep to the company’s portfolio. Hron says these companies come with unique intellectual property that may take too much time or be too costly for Thomson Reuters to build on its own. There are also times where talent acquisition is a key consideration.

“I think the teams that we’ve acquired have been extremely critical—and I would probably include myself in that—in moving Thomson Reuters forward into being a more entrepreneurial and technology-led company,” says Hron.

It’s critical that Thomson Reuters and the smaller companies it acquires be aligned on future revenue and user growth targets, says Hron. But there must also be agreement on what technology integration will look like, especially in the first year after the deal closes. He shares the experience of integrating Materia, a startup that specializes in agentic AI tools for tax, audio, and accounting professionals. 

“Being a small team, you run this risk of swallowing them and their priorities very early on,” says Hron. It would be easy for Thomson Reuters to get excited about Materia’s more-indie tech solutions and quickly plug it into the dozens of tax products the company offers across its portfolio. 

Instead, he advocates that both parties align on as few as three business goals that the Materia team should prioritize. “Everything else that other teams might ask of you are not necessary,” he adds. That allows the smaller startup to remain focused on the specific problems that they were built to solve in the first place.

After the one-year mark, integration accelerates, a reflection of technology maturing and the combined companies having a shared roadmap of future product development, as well as Thomson Reuters wanting to get the most out of the hundreds of millions it may spend when buying these startups. By linking acquired technologies in one product suite, Thomson Reuters can make a more compelling pitch to customers.

“A lot of the value proposition of us acquiring companies is really solving the problem for our customers in a more end-to-end way,” says Hron.

When it comes to integrating and standardizing back-end systems, Hron describes a more pliable process. Since cybersecurity and compliance standards are uniform across all of Thomson Reuters, acquired startups are immediately standardized. But Hron takes a more hands-off approach when it comes to some administrative tools. If a startup’s team prefers to use Slack, they’re allowed to continue to do so even if Thomson Reuters relies on Microsoft Teams. Over time, as those teams blend together and work more collaboratively, everyone will tend to gravitate to the same systems.

Branding decisions are also taken on a case-by-case basis. Materia’s technology was folded into the company’s CoCounsel offering, while the SafeSend and SurePrep remained intact because they are still recognizable brands within the market.

There is also plenty of work that the team has been doing to transform the company’s existing products. In August, Hron says the online research tool Westlaw unveiled Deep Research, an agentic workflow that can plan, review, and strategize legal research in a manner that emulates human work.

With more than 100 products, each with domain specific-functionality, Hron sees many more opportunities to plug in agentic AI to speed up work for legal and tax professionals.

“That really opens the aperture of what is possible for us with our existing portfolio of applications, if we can really expose them in a way that fits within this agentic context that we’re moving towards,” he said.

John Kell

Send thoughts or suggestions to CIO Intelligence here.

NEWS PACKETS

OpenAI, Databricks inks AI agents pact. OpenAI struck a multiyear, $100 million deal to make the AI hyperscaler’s LLMs available for the roughly 20,000 companies that store their data in Databricks’ platform, allowing them to build AI agents with their own enterprise data. The Wall Street Journal reports that Greg Ulrich, chief AI and data officer at Mastercard, said the company is using Databricks to build AI agents in areas such as customer onboarding and support. “The key to getting value out of AI is to be able to link it with your data,” he said. Beyond their new partnership, both OpenAI and Databricks have been ramping up ties to others in the AI ecosystem as annual global spending is projected to ramp to around $1.5 trillion in 2025. 

Starbucks CTO resigns. Deb Hall Lefevre, the CTO at Starbucks since 2022, has reportedly resigned according to Reuters, which has seen an internal memo sent to corporate staff on Monday. Lefevre will be succeeded on an interim basis by Ningyu Chen, previously SVP of global experience technology for the coffee giant. The news comes just after Starbucks approved a $1 billion restructuring plan that will include closing underperforming coffeehouses and job cuts as CEO Brian Niccol aims to turn around results, which includes a pivot away from mobile-only “pickup” stores to recreating a more welcoming, inviting environment in the retail shops. Prior to Starbucks, Lefevre spent 16 years at rival McDonald’s.

Anthropic says its new AI model can code for longer than prior versions. A new model, called Claude Sonnet 4.5, can code on its own for up to 30 hours straight, says Anthropic, a big leap from the prior model called Claude Opus 4 which was said to be able to handle coding tasks for up to seven hours by itself. The new model is also reportedly better at following instructions and intended to be more proficient at using an individual’s computer to take actions for them. As Fortune notes, dueling usage studies published earlier this month from Anthropic and OpenAI found that the former’s products are primarily used for work-related tasks like coding and research, while the latter is used mainly for more personal and exploratory purposes.

ADOPTION CURVE

Marketers say vast majority of agentic AI projects are stuck in the pilot phase. A survey of 200 marketing leaders found that the industry was fairly optimistic about how much AI would change workflows for their industry, even with little progress to show on actual adoption of the technology thus far. More than eight in ten believe that fully automating content creation would reduce “most to all” of their agency spend, with 60% already reporting that they are spending less on agencies in 2025 due to AI, according to a study conducted by generative AI startup Typeface. 

But most of these leaders (82%) say that their agentic AI projects remain stuck in the pilot phase. The top challenges that are preventing wider deployment were concerns around compliance, legal or privacy issues (56%); lack of technical resources or IT support (53%); and poor data quality, which resulted in the AI agent’s output not being useful (48%).

Vishal Sood, chief product officer at Typeface, told Fortune that the company has conversations with CMOs that get super excited about AI’s potential, but that they can be “extremely disconnected from actual day-to-day operators.” He advocates that leaders need to be clearer about their AI strategies and not overwhelm workers with extra work, just because they think AI can handle more asset creation.

Courtesy of Typeface

JOBS RADAR

Hiring:

- Illinois Bone & Joint Institute is seeking a CIO, based in the greater Chicago area. Posted salary range: $260K-$320K/year.

- Glencoe Capital is seeking a head of IT, based in the greater Boston area. Posted salary range: $200K-$225K/year.

- TabaPay is seeking a chief information security officer, based in Palo Alto, California. Posted salary: $250K/year.

- Our Blood Institute is seeking a CIO, based in Oklahoma City. Posted salary range: $200K-$250K/year.

Hired:

- Kratos Defense & Security Solutions has promoted Brian Shepard to the role of CIO, where he will oversee all IT, including security, information compliance, and data communications for the defense industry manufacturer. Shepard has spent 30 years in the IT field, including the last 20 years working at Kratos, most recently as VP of IT.

- Inogen named Naga Rameswamy as CTO, joining the manufacturer of oxygen concentrators from eye care device maker Alcon, where he served as VP and global head of digital health technology. Prior to Alcon, Rameswamy spent nearly 20 years at GE Healthcare, ending his career there as VP of digital health engineering.

- Stella Automotive AI announced the appointment of Fred Seidelman as CTO, where he will drive the technology strategy and product innovation for the provider of automated services and bookings for car dealerships. Seidelman joins from marketing agency McCann Relationship Marketing, where he served as global CTO and oversaw a team of 450.

- Actabl appointed Joseph Benjamin as CTO, joining the hospitality software provider after most recently serving as CTO at recruiting software maker iCIMS, where he oversaw a 430-person team. He also previously served as chief product and technology officer at software maker Behavox and as CTO of DataLogix, which Oracle later acquired for $1.2 billion.

- Ncontracts appointed Sonja Tsiridis as CTO, joining the governance, risk, and compliance software provider after most recently serving as CTO of cybercrime intelligence data provider Intel 471. Tsiridis was also previously CTO of software developer Zoll Data Systems and spent 12 years at software-as-a-service company GHX, where she served as VP of software engineering.

- CreditRiskMonitor.com announced the appointment of Madhav Kale as CTO, joining the credit reporting firm after most recently serving as EVP and CTO at software provider Carixa. He also previously served as CTO of WellAware and as CEO and founder of Zebra Digital Assets.