A lawyer can search case law without LexisNexis, an insurer can underwrite without LexisNexis Risk Solutions, and a researcher can work without Elsevier. A bank can build its own fraud models too.

The more interesting question is what happens after those products become part of the work itself.

A law firm can build research, drafting, and internal knowledge around one legal platform. An insurer can use the same data provider to price policies and investigate claims. A bank can embed fraud and identity checks into onboarding. A scientist may use the same research database every day.

RELX has spent years moving from publishing information to helping customers make decisions with it. The more deeply the product sits inside the decision, the harder it becomes to replace.

RELX still owns journals, databases, and legal content, but that description fits the company less well than it used to.

Electronic products and services were 35% of revenue in 2005. By 2025, they were 84%.

Management has been pushing the portfolio toward analytics and decision tools built around proprietary content and data. The latest trading update suggests that shift is still working.

Underlying revenue grew 7% through the first nine months of 2026. Risk grew 8%, Scientific, Technical & Medical grew 6%, Legal grew 10%, and Exhibitions grew 4%, partly held back by disruption in the Middle East.

Management expects adjusted operating profit to grow faster than revenue in Risk, Scientific, Technical & Medical, and Legal.

Legal is the clearest test of whether artificial intelligence is helping or hurting the franchise. Ten percent underlying growth while AI products are being rolled out is consistent with AI strengthening the offering. The disclosed figures do not isolate how much of that growth came from the new tools.

Risk generated £1.81 billion of first-half revenue and £697 million of adjusted operating profit. The margin was 38.5%.

Its products sit inside fraud, identity, insurance, financial crime, commodity intelligence, government, and other recurring decisions. A bank screens a new customer, an insurer prices a policy, a payment is checked for fraud, and a business screens a counterparty.

The customer is not paying to read a report. It is paying to make a decision quickly and with better information.

RELX combines proprietary data, customer-contributed data, public records, and analytics. A competitor can build another interface. Recreating years of claims, identity, fraud, or transaction history is harder.

Insurance shows the logic well.

Insurers contribute information into shared databases because the combined dataset can identify prior claims, suspicious behavior, pricing variables, and patterns that one insurer would struggle to see on its own.

As participation grows, the dataset becomes more useful. The customer can leave, but it also gives up access to information partly created by the market itself.

Fraud and identity work the same way. A bank sees its own transactions. A network can see patterns across institutions.

A broader dataset makes the product more useful, and deeper use makes it harder to replace.

RELX increasingly sells tools that sit inside a process rather than databases that wait for a user to search them.

A fraud product can score a transaction before approval. An insurance tool can return data during a quote. A legal assistant can research, summarize, draft, and check citations in the same platform.

Replacing that kind of product can require more than retraining employees. The customer may need to rebuild integrations, change internal processes, test the replacement, validate outputs, satisfy compliance requirements, and accept the risk of worse decisions.

Legal generated £959 million of first-half revenue. Underlying revenue growth reached 10%, underlying adjusted operating profit increased 13%, and the adjusted operating margin improved to 21.7%.

Lexis+ with Protégé combines legal research, drafting, analytics, and an AI assistant built around LexisNexis content. Management says Law Firms & Corporate Legal is growing at a double-digit rate as adoption increases.

More than 80% of Legal revenue is subscription-based, and management has said the average subscription length is around three years.

That existing contract base matters. RELX does not need to sell an entirely new platform to an entirely new customer. It can add more useful tools to a relationship that already exists.

The lawyers know the content. The firm already pays for the platform. If the AI product saves time and improves the work, RELX has a path to charge more.

Artificial intelligence is still a real threat.

A lawyer can ask a general model a legal question. A researcher can ask for a paper summary. An underwriter can use an internal model.

That can reduce the value of the old search interface.

The harder-to-replicate assets are underneath it.

A legal answer can sound convincing and still be wrong. A lawyer needs the correct jurisdiction, current law, citations, and source material. A researcher needs the original paper. An insurer needs accurate identity and claims data. A bank needs a fraud score it can defend.

AI is more likely to pressure the interface first. RELX’s stronger assets are the underlying data, permissions, citations, and workflow integrations.

The risk is that someone else owns the interface and RELX becomes a lower-margin data supplier.

Scientific, Technical & Medical generated £1.37 billion of first-half revenue. Underlying revenue grew 6%, underlying adjusted operating profit rose 8%, and adjusted operating profit reached £519 million, a margin of 37.9%.

Elsevier is best known for academic journals, but databases and tools are becoming more important.

Scopus helps researchers track scientific literature and citations. Clinical and education products support medical decisions and training. Products such as Scopus AI, Sherpath AI, and LeapSpace move Elsevier closer to the workflow.

Primary publishing still has real risks. Open access, regulation, library budgets, and criticism of subscription pricing can pressure the traditional model.

RELX does not need journals to disappear. It needs a larger share of the economics to come from tools that help customers use the underlying scientific information.

RELX does not publish a clean price-versus-volume breakdown across the group, so I would not pretend to know exactly how much growth comes from price.

The economics still point toward pricing power in the strongest products.

A large law firm spends heavily on salaries. A research platform that saves lawyers meaningful time does not need to be cheap. An insurer can lose far more from fraud or poor underwriting than it spends on data. A bank can justify paying more for financial-crime tools if they reduce losses or compliance risk.

A publisher can raise price because content is useful. A decision tool can raise price because it saves labor, improves outcomes, or reduces risk.

Much of RELX’s business is subscription and recurring. Legal is more than 80% subscription, and RELX says it has multi-year contracts across much of the broader revenue base.

That means a large amount of revenue is already in place before new products are introduced.

A law firm can move from legal research to legal research plus Protégé. An insurer can add another dataset. A bank can expand fraud tools into another geography.

RELX can sell more value into an existing relationship rather than rebuild distribution from scratch.

Exhibitions generated £575 million of first-half revenue, with 6% underlying growth and a 39.3% adjusted operating margin.

Trade shows depend on physical attendance, exhibitor budgets, travel, and event calendars. They have network effects because exhibitors want buyers and buyers want exhibitors, but the revenue is more cyclical and less embedded than Risk or Legal.

I would not value an event business the same way I value a multi-year fraud or legal subscription.

Exhibitions adds cash generation and diversification. It is not the center of the thesis.

Print and print-related revenue was £157 million in the first half, down from £189 million a year earlier.

That decline has been more than offset by growth in analytics and tools.

Group adjusted operating margin increased from 33.9% in 2024 to 34.8% in 2025 and 35.5% in the first half of 2026. Return on invested capital reached 15.4% in 2025, up from 11.9% in 2021.

Growth has accelerated while margins and returns on capital have improved.

Adjusted cash flow was £1.69 billion in the first half and cash conversion was 98%.

Free cash flow before dividends was £1.16 billion. Capital spending was £292 million, about 6% of revenue, and most of that spending was capitalized product development rather than physical equipment.

RELX does not need factories to grow Risk or Legal. It needs data, software, product development, content, and people.

That leaves substantial cash for dividends, buybacks, acquisitions, and debt service.

RELX spent £1.5 billion on buybacks in 2025. The 2026 authorization increased to £2.25 billion, and £1.75 billion had already been completed by June 30.

The interim dividend increased 7%, and the share count is falling.

Net debt increased to £8.73 billion at the end of June. Net debt to EBITDA was 2.3 times, up from 2.0 times at the end of 2025.

That leverage is manageable given the recurring revenue and strong cash conversion, but it raises the hurdle for aggressive repurchases.

I would still judge buybacks by price. Future repurchases are more attractive at today’s lower valuation than they were when the shares traded materially higher.

RELX spent £270 million across five acquisitions in 2025 and £103 million on two acquisitions in the first half of 2026.

Those numbers are modest relative to the size of the company.

Management generally targets datasets, analytics, and products that fit existing platforms. A fraud dataset can strengthen Risk. A legal technology product can strengthen LexisNexis. The acquired asset can then be distributed through an existing customer base.

I would become more cautious if acquisitions became the main explanation for growth.

So far, organic product development is doing most of the work.

The market is worried that artificial intelligence can make legal research, scientific search, and professional analysis cheaper.

My variant view is narrower.

AI may commoditize parts of the interface without commoditizing the underlying data or the workflow.

RELX’s strongest products combine proprietary or contributory datasets, trusted content, permissions, citations, and integrations inside recurring professional decisions.

The October trading update supports that view so far. Legal underlying revenue grew 10% through the first nine months of 2026, Risk grew 8%, and Scientific, Technical & Medical grew 6%.

Management says AI-enabled analytics and decision tools are contributing to growth rather than cannibalizing it.

I would change my view if AI adoption kept rising while revenue growth, retention, or pricing weakened.

Reinvestment, interest, taxes, and dividends absorb cash before management decides how much to spend on repurchases. In the first half, RELX generated £1.16 billion of free cash flow before dividends and retained £312 million after paying them.

Management completed £1.75 billion of buybacks during that period, so repurchases exceeded internally generated cash after dividends. Net debt ended June at £8.73 billion, or 2.3 times EBITDA.

I would prefer repurchases to remain sensitive to valuation, especially while leverage is higher than it was at the end of 2025.

RELX closed November 2 at £29.98.

Using analyst consensus of roughly 141 pence of adjusted earnings per share for 2026, the stock trades at about 21 times earnings. The 10% EPS growth rate and exit multiples below are my assumptions, not management guidance.

For the 18-month view, I assume earnings compound at 10% annually. That would put earnings per share around 163 pence 18 months from now.

If the multiple stays at 21 times, the implied value is about £34.16. That is roughly 14% price upside, or about 9% annualized before dividends.

At 22 times earnings, the implied value is about £35.79. That is roughly 20% price upside, or about 13% annualized before dividends.

At 23 times earnings, the implied value is about £37.41. That is roughly 25% price upside, or about 16% annualized before dividends.

Assuming approximately £1 per share of dividends over the next 18 months, the 22-times case produces about a 14.6% annualized total return before tax and fees. Future dividends are not guaranteed.

Per-share earnings growth gets me part of the way. A full 15% hurdle probably requires either some rerating, stronger earnings growth, or a lower entry price.

Over three to five years, the durability is easier to underwrite than the exact multiple.

If RELX can compound earnings around 9%–11% while retaining a roughly 20–22 times valuation, the business can still produce a solid long-term return.

The question is whether that is better than the alternatives available in the portfolio.

The strongest counterargument is that RELX’s advantage sits closer to the interface than I think.

If general-purpose AI systems become the primary place lawyers, researchers, insurers, and banks work, those platforms could own the customer relationship and bargain RELX down to a lower-margin data supplier.

That would weaken pricing, distribution, and the valuation multiple.

Legal growth and renewal behavior are the clearest near-term tests.

Net debt at 2.3 times EBITDA is manageable for a business with recurring revenue and 98% cash conversion. The financing risk would become more important if operating growth weakened while management continued using cash aggressively for repurchases.

RELX adds exposure to legal, insurance, fraud, scientific, and professional workflows.

Those earnings drivers are different from exchanges or ratings, but the portfolio would still be adding another information-infrastructure business.

It adds some diversification, but it also increases exposure to the same broad theme: high-margin data and workflow businesses whose value depends on remaining embedded in customer decisions.

I would want a clearly superior expected return before adding that exposure.

Decision: Watch.

Price: £29.98 on November 2, 2026.

I do not own RELX.

RELX has recurring revenue, high margins, strong cash conversion, and improving growth. The October update shows the business is still growing as it rolls out AI products, though it does not isolate their contribution to revenue.

At the current price, my base-case assumptions do not leave enough room between the potential return and the risks. A higher return would require faster earnings growth, a higher valuation multiple, or a lower entry price.

I would become more interested around £26–£27. I would also reconsider the current price if Legal and Risk continue growing near current rates and RELX shows that AI products can support sustained earnings growth without relying on a higher valuation multiple.

I would become less interested if Legal growth slows materially, AI adoption rises without stronger monetization, retention weakens, or third-party AI platforms begin controlling the customer relationship.

Disclosure: I do not own shares of RELX as of the date of publication. This research reflects my personal views and is not investment advice.