Much of the debate about AI in research focuses on technology. We believe the bigger question is commercial. AI is fundamentally reshaping the relationship between agencies and clients in three ways:

1.               The contract between them is changing.
2.               The roles performed are changing.
3.               The value created by research is moving.

The agency-client contract is being rewritten

The AI story began with ELIZA, one of the world’s first chatbots, developed in 1966. Despite being little more than a pattern-matching system, many users believed it genuinely understood them.

Today’s AI wave is different:

• The technology genuinely works
• It’s accessible to almost everyone
• It’s been democratised through tools such as ChatGPT and Claude
• Organisations are seeing measurable value.

This matters because many of the activities traditionally purchased from agencies can now be performed more easily, quicker and at lower cost. Survey design. Coding. Reporting.

The current contract has become obsolete.

As information becomes more abundant, interpretation, challenge and trusted expertise become more valuable. The future agency-client contract will be built less around process and more around expertise, trust and helping organisations navigate the unknown.

Technology is moving faster than organisations

The greatest challenge facing the industry may not be AI itself. It may be organisational adaptability as most are designed for efficiency, governance and stability. Few are designed for continuous reinvention.

Yet AI demands exactly that. AI is not a technology transformation. It is an operating model transformation.

Operating models are built from three interconnected elements: 

• People
• Processes
• Technology

During 10 years leading transformation programmes at Unilever one lesson became clear. The role of transformation is not to react to the future. It is to prepare the organisation for it.

We were experimenting with AI technologies years before generative AI became a mainstream boardroom discussion. More importantly, we continuously evolved our operating model to stay ahead of changing business needs. Across that decade, we redesigned our operating model seven or eight times. Not because it was broken, but because the future had changed.

Many organisations and agencies are not currently structured to evolve at that pace. That may prove to be a bigger risk than AI itself.

Research is not disappearing – its value is moving

Much of the debate focuses on whether AI will replace researchers, shrink agencies or move more research client-side. These are important questions, but they miss the commercial one.

AI will expand the market. More organisations will access research, and existing users will conduct it more frequently. More research will be done, but not all of it will require a specialist. 

Like web design, DIY tools increased participation while squeezing the middle of the market. Research is likely to follow a similar path.

Where does value sit in the new contract? Historically, much of the industry’s revenue has been generated through process execution: survey design, sampling and recruitment, coding and analysis.

AI is rapidly reducing the effort required across these activities with significant implications.

For decades, value was attached to process execution. Increasingly, it will sit in interpretation, judgement, decision intelligence and the ability to drive organisational change and growth.

This shift has significant implications for roles. Some will disappear, others will evolve, and new ones will emerge.

The future operations supplier faces an immediate choice: commoditise or move up the value chain.

The future agency may spend less time executing projects and more time helping clients interpret complexity, and drive action.

The future client-side researcher is likely to spend less time generating insight and more time translating evidence to drive data-driven decisions and growth.

Specialist knowledge was always the real product, but it was never priced as such.

For decades, agencies bundled what they knew into what they did. A specialist in young people didn’t charge for understanding that audience, they charged for the fieldwork, the analysis and the report. The knowledge came with the process. It had no line item of its own because it never needed one.

AI has changed that. Process is now cheap. What remains is what specialist agencies knew all along, and for the first time it has the space to stand alone.

In short, AI does not remove the need for expertise. It increases its importance. As process becomes cheaper and more accessible, expertise, judgement and decision intelligence become the product in their own right.

The future belongs to specialists, not factories. But expertise alone is not enough.

The winners will be those who combine expertise, decision intelligence and the ability to drive organisational change at pace.

Conclusion

The research industry has repeatedly demonstrated its ability to adapt. What makes AI different is not that change is happening. Change has always been a constant.

What is different is the pace at which technology is evolving and the scale of the opportunity it creates.

The traditional agency-client contract was built around process. Agencies sold it. Clients bought it. And for many years that model worked. Today, AI is exposing its limitations.

The real value was never the process itself. It was the expertise, judgement and ability to turn evidence into better decisions.

Process was simply the wrapper. AI has begun to remove it.

Previous AI winters occurred because the technology was not ready. This time it is. The greater risk is that agencies, client-side teams and organisations fail to adapt quickly enough to capture its value.

The next AI winter, if one occurs, is unlikely to be a failure of technology. It may be a failure of relevance. 

Both sides built the existing contract. Both sides will need to rewrite it. Deliberately. Together. With AI embedded in the foundations rather than added as a layer on top.

The contract needs rewriting. The only question is who holds the pen.

Josephine Hansom is a youth researcher and keynote speaker. Alex Owens is an independent adviser.

The full Delphi report, ‘Who owns understanding? How AI is reshaping value, expertise and accountability in research’ is accessible via the MRS website.