Thomson Reuters: 11 insiders bought on July 28, and the timing matters

Thomson Reuters Thomson Reuters Corporation did not wake up on July 28 in a vacuum. The stock had already taken a quick trip lower, from an opening price of $95.25 on July 20 to a close of $87.33 on July 22, then clawed back toward $97.65 by July 27, according to the company’s stock-price history and market quotes. That kind of move forces a fresh read on any insider buying. The chart had already shaken out weak hands and then recovered before the next earnings print.

The July 28 filings came in after that swing, not before it. Karen L. Hirsh bought shares valued at about EUR 5,006, Robert Durnford bought EUR 1,968, Raees Nakhuda bought EUR 4,492 and Caroline Rogge bought EUR 6,032, all on the same day and all flagged as buys. Those are euro-normalised filing values, not local share prices, and they are small in absolute terms against a company with a market value of about EUR 40.47bn. Small does not mean meaningless. It does mean you should read the pattern, not the cheque size.

InsiderTrades data puts the relevant historical bucket, director-level buys at mega-cap names, at a 54.7% 90-day win rate and a 3.95% average return over 90 days. That is historical cohort data, not a forecast for this stock, and it is one reason the cluster matters more than any single ticket.

Why this name matters in the first place

Thomson Reuters sits in specialty business services, but that label undersells the operating tension. The company sells legal, tax, accounting and news information services into a market where technology integration has become the price of admission. In January, Thomson Reuters said the legal industry was going through a tectonic shift, with law firms increasing technology spending by nearly 10 percent and profits per lawyer in the Am Law 100 up 53.7 percent since 2019, driven by AI adoption and client demand shifts. That is the real backdrop here. The company is not trying to sell a generic software story. It is selling workflow tools into a profession that is being pushed, sometimes reluctantly, toward automation.

CoCounsel is central to that story. Thomson Reuters said its AI assistant reached one million users by February 2026. One million users is not a vanity metric in this context. It tells you the company has some distribution in a market where incumbency still counts, and where the buyer is often not shopping for a toy. The company also announced a July 14 joint venture with KKR for the global print business, which tells you management is still pruning legacy exposure while leaning harder into higher-value information products. The stock is not being asked to carry a nostalgia premium. It is being asked to justify a software and workflow multiple in a business that still has old-line assets attached to it.

Peer context helps. RELX operates in overlapping legal and risk analytics segments, and the broader specialty business services group has been trying to prove that AI can lift both product relevance and pricing power. Thomson Reuters is not alone in that effort, but it has a recognizable brand and a deep installed base. That matters when the market is trying to decide whether AI is a feature or a moat. In this corner of the market, the answer is usually both, and neither is free.

July 28, four buys, and the shape of the cluster

The July 28 filings are the hook, but the cluster is the point. Karen L. Hirsh, Robert Durnford, Raees Nakhuda and Caroline Rogge all bought on the same date. InsiderTrades data also shows 11 insiders trading the name in the same direction over the past quarter, with 12 recent declarations in the cluster set. That is not a random trickle. It is a coordinated-looking pattern of buying across director-level names, and our scoring rewards that configuration more than a lone print.

The individual amounts are modest. Hirsh bought EUR 5,006, Durnford EUR 1,968, Nakhuda EUR 4,492 and Rogge EUR 6,032. None of those numbers will move a market cap of EUR 40.47bn. That is not the test. The test is whether a group of insiders chose to add exposure after a sharp pullback and before an earnings date that can reset expectations. On that score, the timing is cleaner than the size.

The insider function attached to the filings is also relevant. These are director or senior officer level buys, not a one-off from a peripheral holder. Our internal signal score for the Hirsh filing came in at 51, and the rationale is straightforward enough: an operating director, part of a wide cluster, and a filing value that is negligible relative to the company. You do not need to overread the score. You do need to notice that the buying is coming from inside the governance layer, not from a passive corner of the cap table.

What the market had already done before the filings hit

The stock’s July path gives the filings some context. A drop from $95.25 on July 20 to $87.33 on July 22 is a fast enough move to change the emotional temperature around a name. By July 27, the shares had rebounded toward $97.65. That sequence matters because insider buying after a decline can look very different from insider buying after a breakout. One is often a response to weakness. The other can be a confirmation of strength. Thomson Reuters was in the first camp, then partially recovered before the filings landed.

That does not make the buys heroic. It makes them legible. If you are a director-level insider and the stock has already bounced back from a short, sharp dip, you are not buying into a panic low. You are buying after the market has had a chance to digest the move and before the next earnings release on August 5. That is a more deliberate choice. It also means the filing is easier to compare with the company’s own operating cadence, because the next quarter will either validate the rebound or expose it as a relief rally.

The broader macro backdrop is not helping or hurting in a simple way. The Federal Reserve held the federal funds rate target range at 3.5 to 3.75 percent on July 29, with a hawkish statement and mixed projections for further 2026 hikes amid inflation concerns and oil price pressure. For a company like Thomson Reuters, that is not a direct trading input the way it would be for a levered cyclical. But it does matter at the margin. Higher-for-longer policy keeps pressure on multiples, and it keeps the market less forgiving of any stumble in a premium information-services name.

August 5 earnings is the next real test

Thomson Reuters said second-quarter 2026 earnings will be released on August 5. That is the next hard date on the calendar, and it is the one that will tell you whether the July rebound had operating support or was just a market bounce. The insider cluster sits right in front of that print, which is why it deserves attention now rather than after the fact.

The company’s fundamental profile is not weak. InsiderTrades data gives it a fundamental score of 60, with a quality score of 80. Those are not a thesis by themselves, and they are not a substitute for the quarter. They do, however, fit a business that has enough recurring demand and enough product relevance to keep attracting capital. In plain English, this is not a distressed balance-sheet story. It is a premium information and workflow story with execution risk around product adoption, pricing and the pace of AI monetisation.

The market is also not short of opinions. Analyst consensus for Thomson Reuters stood at Buy in late July 2026, with some targets cited around $120. That is useful as a reference point, not a verdict. The stock has already moved enough in the last week of July to remind you that consensus targets are not the same thing as a clean path. If the August 5 print shows durable AI traction, stable core demand and no sign that the rebound was just a reflex, the insider buying will look better in hindsight. If the quarter disappoints, the same filings will look like a small, well-timed but ultimately ordinary cluster.

Historical cohort context for director buys

The cohort read is useful because it keeps the filing from being treated as a one-off anecdote. Director-level buys at mega-cap names have historically produced a modest positive average return over 90 days, and the win rate is just over half. That is not a magic number. It is not a promise. It is a reminder that this kind of buying has, on average, been better than random in our dataset, especially when it comes in a cluster rather than as a lone print.

The limits matter just as much. The cohort is a historical bucket, not a live prediction engine. It does not know whether Thomson Reuters will beat on August 5, whether CoCounsel adoption will accelerate again, or whether the market will decide that the recent rebound already priced in the good news. It only tells you that this pattern, in this role and size bucket, has not been noise as often as skeptics like to assume. That is enough to keep the filing on the page. It is not enough to turn it into a trade on its own.

Our scoring lands in the same neighborhood for the same reason. The July 28 filing from Hirsh scored 51, and the cluster structure is the main reason it clears the threshold. The amount is tiny relative to market cap, but the pattern is not. When 11 insiders trade the same name in the same direction over a quarter, you are no longer dealing with a stray personal allocation. You are dealing with repeated internal willingness to add exposure while the market is still deciding how to price the AI and workflow story.

What to watch after the cluster

The first thing to watch is the August 5 earnings release. That is where the company can either reinforce the July rebound or give the market a reason to revisit the $87 handle from July 22. You do not need a grand narrative before then. You need the quarter, the commentary on AI products, and any sign that the legal and professional-services demand backdrop is still translating into revenue quality.

The second thing to watch is whether the buying pattern continues. A single cluster can be meaningful, but repeated declarations would make the signal cleaner. If more director-level names keep buying after the earnings date, the market will have to decide whether this is simple confidence or a more durable internal view on valuation. If the cluster stops here, the July 28 prints still matter, but they will sit more comfortably in the category of opportunistic buying after a dip.

The third thing is the peer tape, especially RELX and other specialty business services names that live in the same broad conversation about AI-enabled information products. Thomson Reuters is not trading in isolation. It is being judged against a sector where investors want evidence that AI is not just a marketing layer. The company has already put one million CoCounsel users on the board and has moved on the print business with KKR. Now it has to show that the operating numbers can keep up with the story.

The insider filings do not settle that question. They do tell you that four named insiders bought on July 28, that 11 insiders have traded the name in the same direction over the past quarter, and that they did so after a sharp July swing and before an August 5 earnings release. That is enough to keep Thomson Reuters on the watchlist, and enough to make the next quarter more interesting than the average information-services print.

This is not investment advice.