Last week Toby Fitzherbert spoke at the inaugural Asymmetrix Summit on how AI is reshaping the way buyers value data businesses and run processes. Here are the views from across his panel on five ways data deals are being impacted in the current market.
1. Resilience is the first question for data businesses
Early-stage diligence on data businesses used to start with questions purely about organic growth, net revenue retention, gross revenue retention, and cash profitability. Those metrics still matter, but they are no longer where diligence starts.
The first questions now are about resilience and a right to exist. What is the data used for, and how defensible is it? How varied are the sources, how far back does the history go, and how much does the product matter to the people using it? And crucially, what is the AI threat from a competitor, an AI-native start-up or customers self-serving?
In a changing market, customers are looking for data, information and insight they can trust – it’s why, despite market volatility, renewal rates are the highest they have been for high-quality businesses. Data businesses that can evidence the distinctiveness and quality of their data or insight offering will be the most sought after.
2. “Proprietary” has to be specific
“Proprietary” turns up in almost every deal process, and the word has been stretched thin. What earns the label today is the complexity of the dataset, not just where it comes from.
Data that is genuinely hard to copy tends to come from many sources, blend human and digital inputs, carry years of history, and be trusted by customers. Human expertise is a big part of that. It used to be something sellers played down over concerns of headcount growth and key person risk, but it is now increasingly a selling point in its own right. Specialists working alongside the technology are what give the data its context, accuracy and timeliness, and that is what customers are paying for.
Incumbency counts too. Where a product sits inside a high-value workflow, customers rarely switch, even when prices rise.
3. AI diligence is part of every workstream
For any deal, AI diligence is now its own workstream alongside commercial, finance or legal diligence – there is no getting away from that. When it comes to AI, there are three pieces of work that actually build conviction.
The first is definitional: separating genuinely proprietary data from enriched public data. The second is a replication test that scores the target’s real outputs head-to-head against models and agentic workflows. Clearly having more proprietary data helps this. The third is competitive analysis split across customers building in-house, AI-native entrants, and what adjacent competitors are doing. For each, we assess how likely it is to materialise, not just whether it’s possible.
4. Sale processes are longer, and front-loaded
Preparing for a sale is now roughly ten times the work it was three years ago. Previously, a seller might have fielded 50 to 100 questions from a buyer at the first-round stage. Today, with buyers running AI across the data room before formal diligence begins, that number can easily hit several hundred, with the top ten surfacing before the first management meeting. Sellers need those answers ready on day one.
The AI questions above also mean that, alongside normal data room documentation, sellers need AI and product documentation, potentially replication test results, customer workflow evidence and a clear narrative on defensibility. The bar for buyers is higher and they are taking more time on their diligence as a result.
For processes, this means everything is taking longer, particularly with strategic buyers who need to bring several internal teams with them, resulting in lengthy exclusivity periods.
5. AI revenue has to earn its valuation
Many data businesses are now building AI products of their own, including MCP (Model Context Protocol) connections that let AI tools plug straight into their data. Putting a value on that revenue is hard when the evidence is still thin, but it is the biggest opportunity in the sector right now.
What helps is a plan that stands up to scrutiny, early results that line up with it (even if they are small), and a CTO or CPO who can tell the story well. CTOs who can talk about the technology in commercial and financial terms make a real difference to valuation. Investors are focused more on AI opening up new opportunities than on simply justifying higher prices, particularly as pricing may shift over time from per seat to per usage. AI features sold separately, with their own revenue and usage figures, should potentially be worth more than ones bundled into the core price.
If you would like to discuss how the data sector is evolving, please get in touch.
