The market research industry’s AI debate has evolved quickly.

Not long ago, the conversation centred on whether researchers should use generative AI at all. Today, that question feels largely settled. AI is already helping teams generate campaign ideas, draft survey questions, create product claims, summarise findings and explore hypotheses at unprecedented speed.

The more important question now is what happens next.

As AI increases the volume of concepts, messages and recommendations being produced, organisations face a growing validation challenge. Generating options has become easier. Determining which of those options will genuinely resonate with real people remains harder.

This shift presents an opportunity for researchers: rather than viewing AI as competition, research teams can position themselves as the critical validation layer between AI-assisted creation and business action.

Speed creates a new quality challenge

The appeal of generative AI is obvious. In seconds, a tool can produce dozens of campaigns, hundreds of ad variations or multiple product positioning routes that might previously have taken days to develop.

Consumers are already embracing these technologies. Kantar’s Global AI Sentiment study found that 82% of people globally have used an AI assistant in the past six months, highlighting just how quickly AI has become part of everyday decision-making. At the same time, trust isn’t universal. Only 40% of people say they would trust an AI agent to handle complex issues for them, suggesting ongoing caution when higher stakes are involved.

A similar tension exists for organisations. Faster content creation does not automatically translate into better decisions.

The reality is that every additional idea generated creates another choice that potentially needs validation. As the volume of AI-generated output increases, the need for efficient human feedback and real consumer evidence becomes even more important.

And for many organisations, the bottleneck now comes down to confidence rather than creation.

AI accelerates exploration, but confidence still comes from evidence

AI-generated outputs should be viewed as a complement to consumer evidence and human expertise, not a substitute for either.

The real advantage lies in combining the speed and scale of AI with validation from real consumers; this helps ensure that promising ideas are both feasible and relevant.

This is particularly important when decisions carry commercial or reputational risk. A product claim that appears credible to an AI model may create confusion among consumers. A marketing message that seems clear internally may be interpreted very differently by the audience it is intended to reach.

Without validation, organisations risk acting on assumptions rather than evidence.

Human feedback reveals what AI may miss

One of the most valuable functions of research is going beyond determining preference. Expert review and real human feedback uncover the nuance, context and unintended meanings that can easily be missed when decisions rely on assumptions alone. Just as importantly, they help identify blind spots that may never have appeared in the original AI-generated output.

Imagine a team that uses AI to quickly generate multiple campaign territories. This produces a broad range of promising creative directions in a fraction of the time of traditional development. But deciding what is most likely to resonate is a different challenge. When tested with real consumers, some territories unexpectedly fail to communicate the intended message.

Or consider a business using AI to draft product claims. Before committing budget to creative production and media placement, researchers test the language with their intended audience. The feedback reveals that some claims are misunderstood, while others fail to feel credible or relevant.

This is where human feedback becomes especially powerful. Consumers can surface ambiguity, unintended interpretations and points of confusion that may never emerge without it.

Perhaps most importantly, they can introduce perspectives researchers and AI systems were not actively looking for.

Researchers should own the validation layer

Researchers are uniquely positioned to help organisations build confidence in important decisions by combining AI-enabled exploration with evidence from real people. Their role is increasingly about applying human judgement, understanding context and identifying potential risks. This aligns with a future in which humans and AI work together.

A practical starting point is to identify which AI-generated outputs could have the greatest consequences if they are wrong. These are often campaign messages, product claims, positioning statements or strategic recommendations that influence significant business decisions.

From there, researchers can introduce rapid feedback loops with real consumers, include unprompted responses that allow unexpected issues to emerge, and document instances where human evidence changes or overturns an AI-generated recommendation.

Doing so creates a record of where validation mattered and where assumptions would have led teams in the wrong direction.

The future belongs to teams that can validate quickly

AI is making it easier than ever to generate ideas. What it cannot do is tell organisations which ideas will resonate in the real world. That responsibility still belongs to researchers and the consumers they represent.

As campaign development accelerates, the competitive advantage will not come from generating more options. It will come from validating them faster and with greater confidence.

AI can generate the possibilities. Human feedback determines which ones deserve to move forward.

• Download How Instant Consumer Feedback Can Prevent Campaign Misfires to learn how rapid human feedback can help keep AI-assisted campaign development grounded in real audience response.

Steve Wigmore is director of agile solutions at Kantar