The conversation around synthetic audiences and AI-generated consumer understanding will undoubtedly be a major theme at this year’s MRS AI Powered Insights Conference 2026. And for good reason. Brands are under pressure to make more decisions than ever before as media fragments and marketing becomes increasingly tailored across channels, audiences and moments.

The pressure is real. According to Harvard Business Review, 74% of executives say the number of decisions they make daily has increased tenfold over the past three years. At the same time, Zappi’s Connected Insights Imperative report shows that data fragmentation has overtaken budget constraints as the number one barrier to effectively using insights.

Against that backdrop, synthetic approaches are generating understandable excitement. But the industry risks focusing on the output rather than the foundation. The most important question is not whether synthetic audiences can replace human research. It is whether organisations have built the connected consumer data assets required to make synthetic insights valuable in the first place.

Synthetic is only as good as the data behind it

The conversation often starts with models. It should start with data.

Synthetic audiences, personas and agents are not magical sources of consumer understanding. They are reflections of the information they are trained on. If that information is incomplete, fragmented, outdated or disconnected from real consumer behaviour, synthetic outputs will inherit those weaknesses.

While "synthetic" has become a catch-all term for approaches ranging from prediction models to digital twins and synthetic respondents, the same principle applies across all of them: the quality of the output depends on the quality of the data beneath it.

This is why data fragmentation matters so much. In our research, 44% of organisations reported hiring for new AI or data integration roles over the past year, yet data fragmentation ( 41%) still overtook budget constraints as the number one barrier to effectively using insights. Organisations are investing heavily in AI, but many are still struggling to connect the consumer data that exists across studies, teams and systems.

The irony is that synthetic approaches become more powerful as consumer intelligence becomes more connected. The organisations most likely to benefit from synthetic are not necessarily those with the most advanced models. They are the ones with continuously refreshed, connected datasets that provide a complete view of consumer behaviour.

The real opportunity is moving from projects to systems

Many discussions about synthetic focus on efficiency. Can research be conducted faster? Can costs be reduced? Can fewer surveys be run?

That is simply the wrong framing. The opportunity for AI is not to conduct less research, but to bring the consumer lens into more decisions.

Historically, research has been concentrated around high-risk decisions: the major product launch, the flagship campaign, the biggest media investment. But as organisations create more concepts, assets, variants and executions, an expanding layer of lower-risk decisions now receives no consumer input at all. Not because marketers don't want it, but because traditional research was never designed to operate at that scale.

This is where synthetic becomes interesting. Not as a replacement for human research, but as a way to expand testing into decisions that would otherwise go unvalidated. It enables organisations to bring consumer input to far more ideas, assets and executions than has historically been practical, helping teams identify which opportunities deserve deeper investment and human validation.

In Zappi’s Connected Insights Framework, this shift represents the move from fragmented ways of working toward connected and ultimately AI-accelerated organisations where insights become embedded in decision-making itself. Within these systems, synthetic approaches become a force multiplier, extending consumer understanding into the many smaller decisions that collectively drive growth.

The future belongs to connected intelligence, not standalone AI

The strongest AI applications are increasingly built on connected systems that combine consumer data, business knowledge, research history and specialist agents. In these environments, the model becomes just one layer in a larger ecosystem.

In innovation, synthetic approaches can help teams pre-screen large volumes of AI-generated concepts before investing in deeper human validation, allowing them to prioritise the ideas with the highest potential.

In advertising, the opportunity is similar. Marketers increasingly create tailored assets and messages across audiences, channels and contexts. Connected systems can provide rapid, consumer-informed feedback throughout that process, helping teams test digital assets and personalised messaging that have historically gone unvalidated – not for lack of interest, but for lack of access.

Synthetic should not be viewed as an alternative to connected insights. It is the next expression of them. Their value is extending consumer understanding into decisions that once relied on instinct, assumptions or internal debate.

___

Steve Phillips is co-founder, executive chair and chief innovation officer at Zappi