Today, AI is reshaping how data businesses sell and monetise what they do. Across our conversations with operators in data and information services, one question comes up again and again: in an AI-first market, what keeps a data business essential? Toby Fitzherbert shares four things he sees market-leading operators doing:
1. Double down on proprietary data, not reasoning
AI is exceptionally good at reasoning over information anyone can access. So, if your product is built on generic, aggregated or publicly sourced data, you’re competing on AI’s home turf.
It helps to picture data as a pyramid:
- The commoditised base: data offered by multiple providers, where customers can switch easily.
- The hard-to-get middle: data that is public, but difficult to access or pull together.
- The proprietary top: data that is expensive to source, has few or no comparables, and is the real reason customers buy.
The volume sits at the bottom, but the value sits at the top. The best data businesses invert their resource allocation accordingly, putting the bulk of their effort into the smallest, highest-value tier.
Data at the top of the pyramid is usually human-driven. Human-to-human relationships remain one of the most important sources of proprietary insight, because AI can’t replicate the trust needed to surface it. Insurance Insider is a great example: its edge comes from deep relationships across the specialty insurance market, and access others simply can’t match.
2. AI is creating new tiers, units and premiums
AI-powered navigation is now table stakes. Most data businesses have moved on, investing in genuine AI products that drive decisions rather than just helping users find things. That shift has a big impact on pricing, and it deserves careful thought.
The best-in-class businesses are approaching pricing in a far more sophisticated way. The first question is the unit: usage-based pricing versus traditional seat-based pricing. Token-based and conversational interfaces are also opening up lower price-point tiers, taking data businesses further down the value chain without cannibalising the core enterprise product.
At the top end, customers are willing to pay significant premiums for unstructured data they can deploy in their own AI systems via an MCP. That should be priced carefully and positioned as a real value add, not simply folded into the existing product suite.
3. Trust and brand are the new battleground
The central tension of the AI era is disintermediation. When your data is consumed inside a third-party LLM or a customer’s internal tool, your brand is one step removed from the end user, and the risk of your data being wrapped in a hallucination rises.
The defence is threefold:
- Source attribution: rigorous attribution, so users always know where an insight came from.
- Governance: robust controls over how data is used and redistributed.
- Owned platforms: continued investment in your own platforms, so the owned experience stays worth returning to. Personalisation on proprietary portals is proving a key retention lever.
Trust is built through structured listening, and that means customer advisory groups, visibly acting on feedback, and beta launches with managed expectations, so an imperfect first version strengthens credibility rather than dents it. And where granular insight is served to non-expert users, guardrails and deep domain knowledge are essential to keep it accurate.
4. Winners rebuild AI-native, they don’t bolt AI on
The most convincing transformation stories come from businesses that have rebuilt their pipeline AI-first, rather than retrofitting AI onto legacy architecture. An AI-native stack lets you swap in new models and techniques as they emerge, while competitors on legacy systems are left re-engineering from scratch each time.
Internally, AI-native pipelines strip out manual collection and collation, freeing subject-matter experts to focus on judgement and the work only they can do. Externally (and this is the bigger prize), they let you build materially better products: more relevant, faster and more closely tailored to each customer’s workflow. That directly increases how much you matter to a client, and how hard you are to replace.
Take Helio Intelligence, the political intelligence provider, which has built a market-leading AI-led monitoring capability. Rather than relying on manual curation of the daily flow of political activity, its AI ingests and structures that content, then surfaces what is genuinely relevant to each client and their policy areas. The result is both prizes at once: more relevant, faster outputs for customers, and more time for its experts to focus on higher-value work.
The common thread across all four is simple. AI rewards the businesses with genuinely proprietary, human-sourced data, and the discipline to rebuild around it.
These are exactly the dynamics we spend our time on in the data and information services sector. If you’d like to discuss your data business, we’d love to talk. Please get in touch.
