Trusted Intelligence: The New Battleground For Research And Data Firms

Over the past 12 months, a quiet but decisive shift has taken place among executives who rely on research, data and expert insight to underpin high-stakes decisions. Many have actively experimented with using generative AI to answer strategic growth, fundraising, competitive positioning, and product development questions. In doing so, they have reached a clear and consistent conclusion: raw AI outputs are not yet decision-grade intelligence.

The limitations are now well understood. Identical queries can generate different answers. Outputs can appear plausible, yet fail under scrutiny. This is because LLMs are probabilistic models: they generate outputs that are statistically likely, not empirically guaranteed. LLM information is harvested from the web, not from a peer-reviewed encyclopedia. Executives know that essential datasets are missing from public models. Most critically, AI-generated responses lack the accountability, context and methodological transparency required for boardroom discussions or investment committee decisions. For business leaders, this is not a marginal technical issue – it is a fundamental trust gap.

The challenge for information services firms – in categories such as technology research, market research, and regulatory content – is that they must rebuild their processes and propositions on a generative AI platform while maintaining the ‘trust premium’ associated with their brand. How can an industry whose product is intrinsically intertwined with trust redefine itself using a technology that is not yet trusted? AI-powered research must be combined with credible experts applying rigorous methodologies to produce consistent, defensible insights. The strength of this model lies not only in the quality of the outputs, but also in the confidence clients place in them.

It is partly because of this challenge that public market sentiment towards information services firms has weakened. Over the past 12 to 18 months, share prices across multiple segments – technology research, market research, and data providers – have fallen to five-year lows, even as broader markets have rallied on the back of AI-driven optimism. Contrast this with the sky-high valuations of LLMs and AI start-ups in private markets. Anthropic, founded in January 2021, recently raised $65 billion in a round that valued the firm at $900 billion. These firms offer the advantages of AI without (for now) the requirement to deliver decision-grade outputs. By contrast, information services firms face either revenue erosion – as some buyers decide that ease outweighs accuracy – or a difficult transition that threatens the core value of their business.

A New Definition Of Trust For Information Providers

It is undeniable that the way enterprise buyers consume intelligence is evolving rapidly. Expectations have shifted from static reports to dynamic, real-time insights. Seat-based pricing models are giving way to enterprise-wide access and usage-based economics. Clients increasingly expect intelligence to be embedded directly into their workflows, rather than accessed through standalone research portals. While AI has not eliminated the need for high-quality, trusted insight, it has certainly raised the bar for speed, accessibility and integration.

For the analyst industry, the question is not whether trust still matters; it is whether incumbent firms can translate that trust into a scalable, AI-native proposition that meets clients’ new expectations. So far, the evidence has not been compelling. Many firms have introduced AI-powered features – chat interfaces, summarization tools, or enhanced search capabilities – but these are often layered onto fundamentally unchanged operating models. From an investor and customer perspective, ‘bolt-on AI’ does little to alter cost structures, accelerate growth, defend against new entrants, expand margins or demonstrate that firms have truly grappled with what quality and credibility mean in the AI age.

Already, over 50% of web content is now thought to be AI-generated, and concerns are mounting that ‘AI slop’ is taking over the information universe. In the same way that bad money drives out good money, the proliferation of low-quality information means that buyers may soon stop placing value on insights or data that cannot be definitively proven to be high-quality. Consequently, research firms adapting to AI-driven expectations will have to go beyond historical approaches to establishing trust, such as brand reputation and analyst credibility.

Both clients and investors will ask new and much more granular questions about the link between trusted intelligence and AI governance. How are models trained and validated? What safeguards are in place to prevent errors or hallucinations? How are human experts involved in validating AI outputs and data inputs? Can insights be explained, traced and audited? Trust is no longer an abstract attribute; nor can it be won through communications initiatives – it is becoming a function of system design.

Winning Business Models Will Be AI-Centric, Not AI-Additive

Against this backdrop, a new model is beginning to emerge: the trusted intelligence provider. This model goes beyond traditional research delivery by integrating proprietary data assets, advanced AI models, expert validation processes, and robust AI governance frameworks into a single, cohesive system. Crucially, it diversifies the point of value delivery from a report, market model, analyst call or workshop into a multitude of AI-enabled interfaces. Intelligence is no longer consumed as a discrete product; it is embedded directly into the client’s decision-making processes and automates work on the client side.

For analyst relations leaders and other research buyers, this transformation has significant implications. The role of analysts is evolving. They are increasingly involved in identifying new datasets to add to AI models, co-designing AI agents with clients, and training small language models to answer role-specific questions. As a result, buyers must rethink how they evaluate and engage with research firms. Traditional metrics such as relevant expertise, methodological breadth and buyer influence remain essential, but they must be complemented by new criteria. Research firms must demonstrate that they have redesigned their operating model around a core AI platform.

For investors with equity in B2B or B2C research firms, the stakes are equally high. The emerging consensus is that neither pure AI platforms nor traditional research firms, in isolation, represent the end state of the market. AI-first entrants lack the historical datasets, industry relationships and depth of trust required for mission-critical decision-making, while legacy firms struggle to achieve the scalability and efficiency that AI enables. The winning model will be a hybrid: industry experts who curate proprietary datasets and deliver value to clients via trustworthy, AI-powered platforms.

Importantly, markets are no longer rewarding AI narratives alone. Announcements of new tools or partnerships have had limited impact on valuations. Instead, investors are looking for tangible evidence of transformation: re-engineered research processes, scalable delivery models, and clear pathways to margin expansion through automation. In this context, the gap between rhetoric and execution has become a central factor in how firms are assessed.

There is also a growing possibility that industry dynamics could shift in unexpected ways. Historically, established players have acquired smaller, innovative firms to extend their capabilities. Current valuation trends raise the prospect of a reversal, where AI-native companies – or investor-backed platforms – acquire legacy assets as a means of accelerating their move into trusted intelligence.

Looking ahead, the most significant changes are still to come. Advances in AI orchestration, data integration, and workflow automation are paving the way for a more embedded model of intelligence delivery. In this future state, insights will surface automatically at the point of decision, powered by AI agents that integrate multiple data sources and operate within enterprise systems. Human expertise will remain essential, but it will be applied differently: focused on collaborative design for client outcomes, data production oversight, and strategic interpretation.

Ultimately, the research and data industry is not moving away from trust; it is redefining it. In the past, trust was largely brand-driven and human-centric, built over time through reputation and relationships. Now there is a third element: trust in a vendor’s AI models, underlying data integrity, governance and policies. For analyst relations professionals and investors, the central question is no longer which firms are trusted today, but which firms are successfully building trusted intelligence platforms – and whether they can scale those platforms fast enough to remain relevant in an AI-driven market.