Nepa was on stage at IIEX Europe 2026 in Amsterdam together with TikTok and Wolt, discussing how AI can help brands understand and improve creative performance on TikTok. The session focused on one of the biggest questions in modern marketing: how do you know what in your creative is actually driving results?
That question also connected closely to the wider conversation at the event.
Across two days, IIEX Europe brought together insights professionals, brands, agencies, technology providers, researchers, strategists and data leaders to explore where consumer understanding is heading next. But this year, the strongest theme was not simply innovation. It was the industry’s move from AI experimentation to AI responsibility.
The question is no longer whether AI can generate summaries, ideas, patterns and recommendations. It is whether those outputs are strong enough to support real business decisions.
That shift matters.
AI has made insight generation faster and more accessible than ever. But when anyone can generate an answer, speed is no longer the competitive advantage. The real advantage is knowing which answers deserve to shape a decision.
For Nepa, IIEX Europe 2026 reflected a wider shift in the industry. AI is no longer just a tool for speeding up research. It is becoming part of how brands understand people, evaluate creative work, optimise marketing and steer growth.
Read the Nepa white paper: Creative AI WhitepaperRead the case: How Creative AI analytics helped Wolt unlock a potential 40%+ CPA reduction on TikTok - Nepa
But the real value of AI will not come from automation alone. It will come from knowing when to use it, when to challenge it and how to connect it to real business outcomes.
IIEX Europe 2026 Executive summary
IIEX Europe 2026 showed an insights industry entering a more demanding phase. The conversation is no longer only about what AI can generate, but about how insight teams can use it responsibly, validate its outputs and turn faster analysis into better business decisions.
A clear theme throughout the event was confidence. AI is making it easier than ever to produce summaries, patterns, ideas and recommendations. But that also creates a new challenge: knowing which insights are strong enough to act on.
As Feranmi Muraina from The Magnum Ice Cream Company highlighted, in a world where anyone can generate an insight, the competitive advantage belongs to the people who know when to trust it, and when not to.
That idea captured the mood of the conference. AI has made insight generation faster and more accessible. But speed alone is no longer the differentiator. The real advantage lies in judgment.
The same theme appeared across several sessions. Guillaume Aimetti from Inspirient and Mark Barraud from JAB Pet Services explored what it takes to fact-check AI-driven quant analysis. Adrien Louis Weinert from Firefish and Anne-Claire Pierron from Unilever Dove focused on the false confidence problem in AI-supported research. Stephanie Douglass from Quest Mindshare and Roddy Knowles from dtect framed market research as the conscience of AI.
Together, these sessions pointed to the same conclusion: the industry is not only trying to use AI more. It is trying to use AI with more discipline.
For brands, this changes the role of insight teams. Their value will not come from producing more reports or faster summaries. It will come from setting clearer standards for evidence, knowing when AI is useful, knowing when human judgment is essential and helping organisations act on insight with greater confidence.
The strongest takeaway from IIEX Europe 2026 is therefore not simply that AI is transforming research. It is that AI is forcing the industry to become more rigorous. The future belongs to insight teams that can combine speed with validation, automation with context and data with the human judgment needed to make better decisions.
The industry has moved from AI excitement to AI responsibility
AI was once again everywhere at IIEX Europe. But the tone has changed.
The first wave of AI adoption in insights was defined by possibility. AI could summarise qualitative interviews, analyse open-ended responses, generate hypotheses, turn data into stories and support faster decision-making. For many teams, the focus was on experimentation: what can we automate, what can we accelerate and what can we do that was not possible before?
In 2026, the conversation became more mature.
The agenda itself showed the shift. Guillaume Aimetti and Mark Barraud’s session, “Trust, But Verify, Fact-Checking AI-Driven Quant Analysis,” addressed the challenge of knowing when to trust AI-supported research outputs. Adrien Louis Weinert and Anne-Claire Pierron’s session, “AI in Research: The False Confidence Problem,” explored how AI can produce expert-looking outputs without necessarily building expert understanding. Stephanie Douglass and Roddy Knowles took the discussion further by positioning market research as a necessary conscience in an AI-shaped knowledge environment.
These were not isolated sessions. They reflected a broader mood at the event: AI is now part of the insights infrastructure, but the industry is still defining the rules for how it should be used. That is the right conversation to have.
AI can make a strong insights function faster, more scalable and more accessible. But it can also make weak evidence look polished. It can produce confident-sounding answers from incomplete inputs. It can create the impression of certainty before the business has earned it.
This is one of the most important risks for insight teams today. The problem is no longer access to insight. The problem is knowing which insight deserves to shape a decision.
Confidence is becoming the new competitive advantage
One of the strongest themes from IIEX Europe 2026 was confidence. This is not just about whether AI tools are accurate. It is about whether the insight function can create confidence in the decisions that follow.
That distinction is important. A business does not need research because it wants more charts, more reports or more dashboards. It needs research because decisions are uncertain, markets are complex and consumer behaviour is difficult to predict.
AI does not remove that uncertainty. In some cases, it makes it easier to hide. The sessions on AI fact-checking, false confidence and the role of research in AI all pointed to the same issue. AI-generated answers can sound clear even when the evidence is weak. They can summarise a pattern without understanding the business context. They can produce recommendations without understanding the commercial risk behind the decision.
For insight teams, this creates a new type of responsibility. They need to define what good evidence looks like. They need to know when an AI-generated output is useful as a starting point, and when it needs human review, additional data or real-world validation. They need to help stakeholders understand not only what the data says, but how confident the business should be before acting on it.
A quick AI-generated read may be enough for early ideation. It may help shape a workshop, challenge internal assumptions or narrow down options. But it is not always enough to support a product launch, a major media investment or a strategic brand decision.
This is why Feranmi Muraina’s point felt so relevant to the wider conference conversation. In a world where anyone can generate an insight, the competitive advantage belongs to the people who know when to trust it, and when not to.
The future of AI in insights will therefore depend less on access to tools and more on the discipline around them.
False confidence is the new research risk
The most dangerous insight is not always the wrong one. It is the one that sounds right too early. That is why false confidence was such an important theme at IIEX Europe 2026.
Adrien Louis Weinert from Firefish and Anne-Claire Pierron from Unilever Dove addressed this directly in their session on AI in research. Their focus reflected one of the biggest risks facing the industry: AI is putting research and analysis capabilities into more hands than ever, but not everyone using those capabilities will know how to judge the quality behind the output.
This is both an opportunity and a challenge.
- On one hand, AI can make insight more democratic. It can help more people ask questions, explore data and understand consumers. It can reduce manual work and make knowledge easier to access across an organisation.
- On the other hand, it can create expert-looking outputs without expert interpretation.
That matters because business decisions are often made under pressure. Stakeholders want speed. They want clarity. They want an answer. AI can give them one. But that answer may not always deserve the confidence it creates. This is where the role of the insights professional becomes more important, not less.
The value of a researcher is no longer only in producing the analysis. It is in knowing what sits behind it. Was the question framed correctly? Was the sample relevant? Are the assumptions reasonable? Is the evidence strong enough? What alternative explanations exist? What should not be concluded from this data?
AI can help answer questions. But insight professionals help judge whether the answer is good enough. That judgment is becoming one of the most valuable skills in the industry.
AI needs a foundation, not just a prompt
Another important message from IIEX Europe 2026 was that AI does not work in isolation. Alex Dobromir from Product Hub by MMR made this point in the session “The Foundation That Makes AI Safe and Insights Fast.” The framing was clear: AI did not transform their research platform by itself. The foundation came from years of structured data, standardised workflows and deterministic automation.
That is an important lesson for any brand looking to scale AI in insights. It is easy to talk about AI as if the main challenge is choosing the right model, writing better prompts or finding the right interface. But the practical truth is that AI is only as useful as the system beneath it.
AI does not magically fix fragmented data, unclear KPIs or inconsistent research design. In fact, it can expose those weaknesses. If previous studies are difficult to compare, AI will struggle to connect them meaningfully. If brand metrics are not clearly defined, AI can summarise them without making them more useful. If data sources sit in separate systems, AI may make access easier, but not necessarily create better decisions.
For brands, this is a crucial point. The organisations that benefit most from AI will not necessarily be the ones using the most advanced tools. They will be the ones with the strongest insight foundations: clean data, consistent methods, clear taxonomies, connected knowledge and shared decision frameworks.
Without that foundation, AI risks becoming another layer of noise. With the right foundation, AI can help insight teams move faster without losing control.
From insight generation to decision impact
Another major theme at IIEX Europe 2026 was the need to move from insight production to decision impact. Jane Reynolds from Stravito and Laurence Minisini from Givaudan explored this in the session “Why Good Data Still Fails to Shape Decisions, and How Givaudan Fixed It.” The session focused on a problem many organisations recognise: they often have the insights they need, but struggle to use them consistently at the moment of decision.
That is the real challenge. Many organisations already have more data than they can use. They have dashboards, trackers, campaign reports, ad hoc studies, customer feedback, social data and performance metrics.
The issue is rarely a lack of information.
The issue is that insight does not always reach the right people, in the right format, at the right moment, with enough clarity to influence what happens next.
Other sessions reinforced the same point from different angles. Nadine van Rooyen and Irina Ene from WEX Inc challenged how the industry defines insight impact. Patrick Young and Rosie Paul from PRS IN VIVO asked a sharper question in their “Right-First-Time” session: in a world chasing faster and cheaper research, what actually delivers better decisions?
That question should sit at the centre of the industry’s AI conversation. The future of insights cannot only be about generating more outputs. It has to be about improving the quality of decisions.
For insight teams, this means moving closer to the decision process. It is not enough to answer research questions. The goal is to help the business understand the implications, weigh the evidence and choose the right next step.
This matters especially for brand tracking, campaign measurement and marketing effectiveness work. These programmes can generate large volumes of data over time. But their value does not come from the volume of reporting. It comes from helping teams decide what to do differently.
The next competitive advantage in insights will not be speed alone. It will be confidence.
Human judgment is becoming more visible
One of the clearest conclusions from IIEX Europe 2026 was that AI has not made researchers less important. It has made their judgment more visible.
Petra Langen from The Lifesights Company and Patrick Mönkemöller from Ritter Sport captured this well in “AI Tasting Spoon: AI Finds Patterns. Humans Find Meaning.” Their framing gets to the heart of the issue. AI can identify patterns, but humans still need to decide what matters.
Athena Chen from STRAT7 and Luke Hand from Mail Metro Media also pushed the conversation in a human-first direction, exploring the future of insight beyond the dominant tech and AI narrative. Jenny Lindsay from buzzback and Miriam Schwarz from Nestlé made a similar point in the context of early-stage innovation: becoming a better innovator requires more than speed. Together, these sessions challenged a simplistic view of AI.
AI can process patterns, summarise data, support ideation and help teams work faster. But it does not automatically understand context, nuance, uncertainty or commercial consequences. That is where human expertise still matters.
Insight is rarely just about finding an answer. It is about understanding what the answer means. It is about knowing how confident to be. It is about deciding whether the business should act, wait, test again or look at the problem differently.
AI can summarise what people say. But it does not always understand what they feel, avoid, misremember or struggle to express. That is why human-centred methods remain essential. Qualitative research, communities, cultural understanding, video intelligence, behavioural testing and in-context validation all help keep research grounded in real human behaviour.
The strongest teams will not choose between AI and human insight. They will build systems where each does what it does best. AI can help with speed, scale and structure.
Humans are still needed for judgment, context and interpretation.
Synthetic data and AI personas need clear boundaries
Synthetic data, digital twins and AI personas were also central to the IIEX Europe 2026 conversation.
Adam Bai from Panoplai led a “Digital Twin & Synthetic Masterclass” focused on validation and adoption. That framing is important. The most useful discussion around synthetic data is not just what it can generate, but how it should be validated and where its limits sit.
Joseph Rini from Market Logic Software and Sehnaz Arasan from Philips also brought AI personas into the innovation conversation, showing how a campaign idea can meet its first “consumer” before it reaches a traditional research process. Nelly Mamyan from Yasna AI and Anh-Thi Mai from Euroconsumers explored how AI-powered approaches can support faster idea validation across markets.
These tools can be useful, especially early in the innovation process. They can help teams explore hypotheses, pressure-test ideas, identify possible objections and move faster before investing in larger studies. That can be valuable. It can make research more accessible. It can help teams avoid obvious mistakes earlier. It can support faster iteration across product development, communications and innovation.
But the strongest discussions around synthetic approaches were not only about what the tools can do. They were about validation. That distinction matters.
Synthetic respondents should not be treated as a full replacement for real consumers. AI personas should not become a substitute for real-world evidence. Digital twins can support exploration, but their quality depends on the data, assumptions and models behind them.
The best use case is not AI instead of research. It is AI as an early-stage tool that helps teams ask better questions, move faster and decide what deserves deeper validation. Synthetic approaches can be useful for direction. But they need clear boundaries before they are used for decisions with commercial risk.
The say-do gap is still one of research’s biggest challenges
Another clear theme at IIEX Europe 2026 was the gap between what people say and what they actually do.
Daniel Putsche from Horizon and Stephanie Rabbe from Miele focused on this directly in “Miele: From Say-Do Gap to Behavioural Consumer Centricity.” Their session looked at how pretotyping and behavioural experimentation can help organisations move beyond stated intention and closer to real consumer behaviour.
Other sessions reinforced the same theme. Israel Ogunseye from PawaTech explored how behavioural data can unlock fintech adoption in Africa. Paul van Gendt from i-Genie.ai and Mujde Bulut from Danone challenged the idea that surveys can always tell the full truth. Tamanna Dhamija and Tim Carey from Convotrack showed how video intelligence can uncover hidden category truths by observing what consumers do, not only what they say.
This is not a new challenge in research. But it is becoming more important as AI makes it easier to generate fast answers from stated data.
Surveys and self-reported responses remain valuable. They help brands understand attitudes, perceptions, motivations and preferences. But they do not always capture real behaviour, trade-offs or decision context.
That is why behavioural validation remains so important. People may say they like a concept, remember a brand or intend to buy a product. But what they actually do in a store, on a platform or in a real decision moment can be different.
For brands, this matters because many commercial decisions depend on behaviour, not only stated preference. The future of research will not be about choosing between stated and behavioural data. It will be about combining them more intelligently. The strongest insight systems will connect what people say, what they feel, what they notice and what they actually do.
Brand tracking must become a decision tool
Brand tracking was another relevant theme at IIEX Europe 2026. But the role of tracking is changing.
Elisabeth Schinke from quantilope and Frank Hofmann from Savencia explored this in their session on Savencia’s strategy for brand tracking using mental availability. Their focus reflected a broader shift: brand tracking is no longer only about following traditional funnel metrics. It is increasingly about understanding the mental structures that support brand growth.
Traditional brand tracking has often focused on monitoring: awareness, consideration, preference, image, loyalty and other key brand metrics. These remain important. But modern marketing teams need more than passive measurement.
They need to understand what is changing, what is driving the change, what it means for growth and what action to take next. That requires brand tracking to become more connected to business outcomes. It also requires better storytelling, clearer frameworks and more frequent activation.
A tracker should not just tell a brand team whether the numbers are moving. It should help them understand why they are moving, what the movement means and how to respond. This is also how we see the future of brand tracking at Nepa: not as a passive reporting system, but as a decision tool for brand, media, creative and commercial teams.
The most future-ready tracking systems are becoming more modular, more connected to campaign and media data, and more useful across functions. They serve as a shared source of truth for the business, not just a reporting tool for the insights team.
In a more complex marketing environment, that matters. Brands do not only need to know how they are performing. They need to know what to do next.
Creative performance is becoming more measurable
Nepa’s session with TikTok and Wolt focused on one of the most important questions in marketing today: how do you know what in your creative is actually driving performance?
Marketers often know which assets performed well. But they do not always know why. Was it the opening second? The product shot? The format? The message? The creator? The pace? The emotional tone? The call to action?
This is where AI can become genuinely useful.
The wider agenda showed that this topic was not only relevant to Nepa. Thomas Zoëga Ramsøy from Neurons explored how AI can help brands predict, optimise and visualise advertising before launch. Wim Hamaekers from One Inch Whale and Uyen Chand from Bose showed how AI, colour science and human insight can be combined to optimise colour and drive consumer preference. Katie O’Connor from Behaviorally and Cesar Lastra from Royal Canin discussed the adoption of AI-powered pack testing inside global brands.
Together, these sessions pointed to a broader shift: creative, pack and product evaluation are becoming more measurable, more predictive and more connected to performance.
By analysing creative elements at scale and connecting them to performance KPIs, AI can help teams understand what actually drives results. The goal is not to replace creative intuition. It is to give marketers a clearer feedback loop, so they can make better creative decisions over time.
That distinction is important. Creative performance is not only about measuring winners and losers after a campaign has run. It is about learning which creative choices drive attention, emotion, brand impact and action, then using that learning to improve the next execution.
This is especially important in fast-moving platforms such as TikTok, where creative performance can depend on details that are easy to miss: the first few seconds, pacing, framing, product presence, sound, format and emotional tone.
AI will not replace creative judgment. But it can make creative learning more systematic.
Product and innovation research is moving earlier in the process
Another theme across IIEX Europe 2026 was the movement of insight earlier into the innovation process.
Carolina Toussaint from Unilever and Lisa Nel from Conveo explored AI-led design thinking for product innovation. Their session showed how AI-led research can support design thinking at scale, without losing depth. Joseph Rini and Sehnaz Arasan’s Philips session also pointed in this direction, showing how AI personas can help campaign or product ideas meet an early consumer perspective before a traditional research process begins.
Niamh McCarthy from Pentland Brands and William Rowntree from SMG brought a different angle, showing how communities can help build better products by bringing consumer voice into the product process. Kati LaBeaume from Oura also connected AI-ready innovation to the importance of a human context layer.
The message across these sessions was clear: insight is becoming more embedded in product and innovation workflows. Rather than waiting until the end of the process to test whether something works, brands are using insight earlier to shape, challenge and improve ideas while they are still flexible.
This is where AI can be genuinely useful. It can help teams move faster, explore more directions and pressure-test assumptions earlier. But the same rule applies: speed is only useful if it improves the decision.
The strongest innovation systems will not use AI to skip consumer understanding. They will use AI to ask better questions earlier, then combine it with real human feedback, behavioural validation and commercial judgment.
The future of research is a smarter mix of methods
One of the clearest takeaways from IIEX Europe 2026 was that the future of research is not one method.
Traditional surveys are still important. But they are increasingly being combined with AI-assisted analysis, behavioural data, video intelligence, digital twins, synthetic respondents, insight communities, product experiments and real-world validation.
This came through across many different parts of the agenda. Monika Karamchandani from Mondelez highlighted the value of projective and gamified techniques in helping consumers express what they may not say directly. Kati LaBeaume from Oura focused on insight communities as a human context layer for AI-ready innovation. Paul van Gendt and Mujde Bulut challenged the limits of surveys. Tamanna Dhamija and Tim Carey showed how observed behaviour can uncover category truths that conventional research may miss.
The strongest research teams are not choosing between old and new methods. They are learning when each method is useful, what level of risk it can support and how different sources of evidence can work together.
This is a more mature way to think about innovation in research. The question is not whether a method is traditional or new. The question is what decision it supports, how reliable it is and what evidence is missing.
That mindset will become increasingly important as research teams are asked to move faster without lowering quality. A smarter research mix is not about adding more methods for the sake of it. It is about matching the method to the decision. Some questions need speed. Some need depth. Some need behavioural proof. Some need long-term tracking. Some need creative diagnostics. Some need a combination of all of these. The best insight teams will know the difference.
What this means for brands and insight teams
The biggest takeaway from IIEX Europe 2026 is that the insights industry is entering a more demanding phase.
It is no longer enough to experiment with AI. It is no longer enough to generate faster outputs. And it is no longer enough to produce more data.
The next competitive advantage will come from building insight systems that are faster, more trusted, more connected and more decision-oriented.
- For client-side insight teams, this means becoming stronger at setting standards. Teams need to define what good evidence looks like, when AI can be used, when human review is required and how uncertainty should be communicated.
- For marketers, it means using AI to create better feedback loops across creative, media, brand and business outcomes. The goal is not just more efficient analysis, but better decisions.
- For agencies and research partners, it means moving beyond speed as the main selling point. The real value is helping clients interpret, validate and activate insights in ways that drive growth.
- For leadership teams, it means recognising that AI does not remove the need for insight expertise. It makes that expertise more valuable. AI has made insight generation easier. But it has also made trusted decision-making harder.
That is where insight teams now prove their value.
The future belongs to teams that know what to trust
IIEX Europe 2026 showed an industry that is no longer debating whether AI matters. That question has been answered. The more important question is how to use AI well.
AI can make research faster. It can make analysis more scalable. It can make storytelling more efficient. But it cannot replace the human judgment needed to understand context, assess risk and make better business decisions.The future will not belong to the teams that generate the most answers. It will belong to the teams that know which answers to trust, how to validate them and how to turn them into decisions that drive growth.
For Nepa, that is exactly where the future of insights is heading: towards smarter systems, stronger evidence and clearer decisions that help brands grow.
Questions or reflections on the trends from IIEX Europe? Get in touch with us.
Published on: 26TH JUN 2026