Ones to Watch: The 5 AI-Driven Insights Agencies Rewriting the Rules

The insights industry is in the middle of a genuine disruption, and a new cohort of agencies is demonstrating that the future belongs to those who move fastest, think most clearly about what AI is for, but don't lose sight of the human truth at the centre of all good research.

Here are 5 agencies I'm genuinely excited about. They are building proprietary AI tools, questioning received wisdom about synthetic data and digital twins, and winning recognition from the most demanding client-side researchers in the world. They sit at a different point on the risk curve from the larger incumbents, which means they can experiment quickly, embed AI structurally instead of bolting it on, and attract talent that wants to build rather than maintain.

Focaldata: the hybrid research partner that punches well above its weight

If you were designing an insights agency from scratch for the AI era, you would probably end up with something close to Focaldata. They combine a tech platform with the kind of expert rigour you would expect from a consultancy, with the ability to handle ambitious global programmes.

This London-based agency has built a genuinely hybrid model: AI for survey design, data collection, qualitative moderation, and initial analysis; with expert researchers - a mix of PhDs, polling specialists, veterans of the most rigorous academic and commercial traditions - taking over where judgment, strategy, and nuance are required. AI makes the research faster and more scalable than a traditional agency could ever be; the human expertise makes the outputs more trustworthy and strategically relevant than a pure-tech platform.

They also boast an impressive client list. For an agency of Focaldata's size, winning Virgin Media O2, Nestle, AB InBev, TikTok, Starbucks, the BBC, and PwC alongside the Rockefeller Foundation, the Charities Aid Foundation, and over forty political campaigns, represents a span of methodological credibility rare at any scale. They have surveyed 56,000 individuals in 101 countries for a single client, fielded 36,000-person global opinion research programmes in days rather than months, and delivered weekly multi-market brand tracking for media agencies - production-grade research at a scale and speed that challenges the traditional value proposition of much larger competitors.

Their AI Qual platform is the product that most clearly articulates the disruptive promise. The ability to conduct genuinely deep qualitative research - the kind that surfaces motivation, context, and nuance, not just stated preference - across hundreds of respondents in hours rather than weeks changes the economics and timelines of qual fundamentally. At 2.7 million AI qualitative interview minutes, they are truly operating at scale.

What keeps Focaldata grounded is a commitment to data quality. Their seven-layer automated quality framework removes up to 35% of sample responses before a single insight is delivered - a deliberate signal that speed and quality are not in tension if the infrastructure is built correctly.

Verve: where human truth is the infrastructure, not the afterthought

In the current rush to synthetic respondents, AI-generated personas, and data-at-speed, Verve is making a pointed and increasingly influential argument: the problem is not that AI is being used, it is that AI is being asked to work without adequate human foundations. Feed a large language model low-quality, generic training data and you get synthetic insights that are fast, scalable, and fundamentally unreliable. But feed it rich, proprietary, contextually deep human data, and something genuinely transformative becomes possible.

This is the logic behind Verve's three-pillar model: Artificial Intelligence, Human Intelligence, and Cultural Intelligence working in deliberate combination. Their flagship product, Verve Intelligent Personas (VIPs), was voted AURA's Innovation of the Year 2024 by client-side researchers [1]. VIPs are AI-powered consumer simulations built on Verve's proprietary community data, not on publicly available datasets. The simulations are grounded in genuine human responses; the AI amplifies, interrogates, and extrapolates; the human researchers apply the cultural context that neither the data nor the model can provide unaided.

The Cultural Intelligence strand is perhaps the most underappreciated differentiator. Verve won the MRS Virginia Valentine Award for Cultural Insights - recognition from the industry's most rigorous professional body that their work in semiotics, trends, and cultural analysis achieves something that data science alone cannot. Their in-house cultural experts decode culture at scale, curating datasets that enrich AI audience simulations with the kind of contextual texture that makes the difference between insight and information.

Their current positioning, articulated by founder and CEO Andrew Cooper in published thought leadership, sharpens the competitive argument further. Cooper argues that 'simulation, not synthetic, will define what's next' [2]: that the industry's enthusiasm for synthetic respondents is generating a lot of heat and relatively little light, and that the more productive question is how to build AI simulations that are genuinely grounded in human truth. It is a provocation aimed squarely at the wave of synthetic-first platforms, and it is landing at the right moment.

At ESOMAR 2025, Verve took home two awards: a Gold for Outstanding Contribution to Consumer Insights with Samsung (Project POP), and a Silver for Excellence in AI and Automation in Market Research with Mars Petcare (Empathy at Scale) [3]. Two podium finishes with two of the world's most demanding brand clients adds the kind of external validation that makes the argument very hard to dismiss.

Day One Strategy: AI built for healthcare, by people who know healthcare

There is a version of AI adoption in market research that looks impressive on paper but collapses under scrutiny - generic language models applied to proprietary research challenges, producing outputs that are fast but fundamentally shallow. Day One Strategy has taken a deliberately different path.

Founded in 2019 by Hannah Mann and Abigail Stuart, making it one of the very few 100% women-owned and managed agencies in the healthcare insights space, Day One has spent five years building an AI capability that is specific, rigorous, and grounded in the particular demands of pharma and healthcare research. Their thesis is simple: AI should not replace the expertise of researchers who understand complex disease areas like oncology, immunology, or rare disease; it should make that expertise work harder and faster.

The result is a suite of proprietary tools that are genuinely distinctive. FutureScope is a tech-enabled approach to value proposition development, enabling pharmaceutical clients to test and stress-test propositions for assets still years from launch. CreAItive Loops is an iterative AI-powered learning system for communications optimisation - rather than running a single ad testing study and drawing a line, it creates a continuous loop of refinement that makes each iteration smarter than the last. AgileTrack 360 delivers weekly brand tracking with simplified data integration, giving clients the kind of always-on brand intelligence previously only available to the largest budgets. Their recently launched solution InsightBrain offers AI-powered precision intelligence, combining deep human expertise with purpose-built technology to surface critical insights from fragmented data [4].

What is particularly sharp about Day One's positioning is their willingness to engage openly with the most contested questions in healthcare AI - including synthetic respondents. Their published thinking on this topic has been cited in specialist pharma media, and their approach (cautious, evidence-led, aware of the specific regulatory stakes in their sector) reflects the maturity of researchers who have been living with these tools rather than demonstrating them.

Forbes Council articles and the BHBIA Best of British Intelligence (BOBI) Award for Best Use of Innovation [5] are the kind of external validation that matters in a sector where credibility is everything. In 2025, they were also shortlisted for an ESOMAR Award for their work with Sanofi on AI-based synthetic personas for communications research [6]. Day One was building around AI before that phrase became a marketing cliche.

Bolt Insight: the qual-at-scale platform built by researchers

There is something refreshing about a company whose founding story is not a technologist discovering an industry, but an industry practitioner deciding to fix it from the inside. Hakan Yurdakul launched Bolt Insight in 2019 having worked client-side at one of the world's largest FMCG companies, where he experienced firsthand how difficult it was to get deep qualitative insight quickly. Research was too slow, too expensive, and too rigid to support the pace of modern commercial decision-making. So BoltChatAI was built to solve the problem he had actually lived.

The platform uses AI moderation to conduct real-time consumer conversations at scale - across 50+ markets, in multiple formats, capturing text, video, and audio responses simultaneously. Clients routinely gather hundreds of in-depth responses within days of briefing, at a fraction of the cost and time of traditional qualitative fieldwork. The AI probing is genuinely dynamic - it does not follow a script, but responds to what respondents actually say, pursuing threads that a human moderator might pursue and surfacing the kinds of motivated, contextual responses that distinguish good qualitative research from survey data in disguise.

The most recent product additions reveal where the platform is heading. Dynamic Personas are AI-powered audience profiles built from real research data - surveys, interviews, trend reports, social listening - that evolve as new data is added. Unlike the static segment profiles that most research agencies still deliver, Dynamic Personas are living representations of audiences that get smarter with every new data point. Meta-Analysis removes the manual burden of synthesising across past studies, allowing clients to interrogate their entire research history in minutes rather than weeks and surface trends, behavioural shifts, and emerging patterns that would otherwise remain buried in presentation decks.

Winners of Best Online Qual at the 2025 MRS Operations Awards [7] and shortlisted by ESOMAR in 2025 for their work with Bahlsen - revealing 'the secret life of cravings' through AI-moderated conversations [8] - they are making their mark. In December 2025, they secured a 7 million pound investment round led by Pembroke VCT [9], with a stated ambition to build what they describe as the 'Assistant AI Chief Insights Officer': a real-time intelligence system that can answer critical business questions instantly, drawing on a brand's entire historical research base as well as live consumer conversations.

Bolt Insight's ambition is to shift insights from reactive to continuous - from commissioning a new study for every question to building an always-on intelligence capability that improves with use. That ambition, if realised at scale, represents a fundamental change in the business model of qualitative research.

Human8: the human spark in the age of automation

As AI becomes the default infrastructure of marketing, a new competitive risk is emerging for brands: distinctiveness erosion. If every brand is using the same AI models to generate the same category insights, develop the same consumer personas, and test the same creative territories, the result is a market full of well-optimised products and communications that are indistinguishable from each other. Human8 was formed, in part, to address exactly this problem.

The consultancy, which describes itself as 'human-driven' with deliberate emphasis, is the result of ten insights businesses combining their capabilities over five years, bringing together expertise in online communities, behavioural science, cultural intelligence, digital ethnography, and AI-led innovation in a single integrated offer. Their CEO Tim Wragg's framing is precise: 'Brands are increasingly turning to AI-driven marketing performance to boost efficiency, often at the risk of losing distinctiveness and authenticity' [10]. Human8's offer is built around preserving and amplifying what they call the 'human spark', not replacing it.

Their AI approach reflects this philosophy at every level. Their AI-canvas workshops - which helped Philips accelerate product innovation while keeping real consumer needs at the centre of every decision - combine qualitative insight from online communities with proprietary AI tools in a way that makes human researchers more productive rather than redundant. The AI Market Pulse tool analyses competitor product launches to detect emerging market trends and reverse-engineer problem statements, giving innovation teams strategic foresight grounded in real market dynamics.

In October 2025 they appointed a new Chief Innovation Officer, Annelies Verhaeghe, focused explicitly on developing human-based digital twins - AI simulations sourced from Human8's own proprietary research communities [11]. These are not simulations built on publicly available internet data; they are simulations built on the richest possible human foundations.

The agency received three international award nominations in 2025 - two ESOMAR shortlistings (for work with Philips and Mars) and one Quirk's Healthcare Research Project Award nomination (for work with Haleon) [12] - reflecting the breadth and ambition of their methodology. The Mars project brought genuine emotion into snacking strategy at a global scale. The Haleon project required deep empathetic research into erectile dysfunction to unlock bold, human-centred messaging for a pioneering brand launch.

What these five agencies tell us about where the industry is going

What is striking about Focaldata, Verve, Day One Strategy, Bolt Insight, and Human8 is not just their individual achievements, but the pattern they collectively represent. AI has been adopted as a foundation rather than a feature. It has been designed to be specific to their methodologies, their client challenges, and their views on what good insight actually requires. That distinctiveness - proprietary data, proprietary tools, proprietary thinking - is becoming the central competitive asset in insights, as generic AI makes the commodity end of the market less valuable and the differentiated end more valuable.

All five also share a commitment to human intelligence that resists the industry's drift towards full automation. They understand what AI cannot do - at least not yet, and perhaps not ever - when it comes to cultural context, empathetic interpretation, and the kind of strategic judgment that turns data into decisions. The agencies that will lead the next decade of insights are not the ones that eliminate researchers the fastest; they are the ones that figure out most clearly what researchers are for in an AI-enabled world.

The great thing is the agencies identified on this list are small enough to move fast but also focused enough to move well. That combination, during a moment of real industry disruption, is what makes them worth paying attention to.

- - -

Sources

[1] Verve Intelligent Personas (VIPs) - AURA Innovation of the Year 2024 - https://www.addverve.com/inspiration/verve-intelligent-personas-voted-innovation-of-the-year-award/

[2] Andrew Cooper, Verve - 'Why simulation, not synthetic, will define what's next' - https://www.addverve.com/inspiration/why-simulation-not-synthetic-will-define-whats-next/ Also referenced in MRS Research Live Industry Report 2025

[3] ESOMAR 2025 Award winners - Gold (Samsung/Verve) and Silver (Mars/Verve) - https://www.wcia.com/business/press-releases/ein-presswire/854015926/winners-of-the-2025-esomar-awards-are-announced-in-prague/

[4] Day One Strategy - proprietary tools including InsightBrain - https://www.dayonestrategy.com/

[5] Day One Strategy - BHBIA BOBI Award for Best Use of Innovation - https://www.dayonestrategy.com/our-thinking/bhbia-bobi-best-of-british-intelligence-award-winner-best-use-of-innovation/

[6] ESOMAR 2025 finalists - Sanofi and Day One Strategy - https://esomar.org/newsroom/2025-esomar-award-finalists

[7] MRS Operations Awards 2025 (Oppies) - Best Online Qual: Bolt Insight - https://www.research-live.com/article/news/oppies-winners-named-for-2025/id/5143333

[8] ESOMAR 2025 finalists - BoltChatAI and Bahlsen 'The Secret Life of Cravings' - https://esomar.org/newsroom/2025-esomar-award-finalists

[9] Bolt Insight secures 7 million pound funding round - December 2025 - https://www.boltinsight.com/bolt-insight-receives-7m-in-funding/

[10] Tim Wragg, CEO Human8 - quote from Chief Innovation Officer appointment announcement - https://www.wearehuman8.com/news/human8-names-chief-innovation-officer/

[11] Human8 appoints Annelies Verhaeghe as Chief Innovation Officer - October 2025 - https://www.wearehuman8.com/news/human8-names-chief-innovation-officer/

[12] Human8 named finalist in three global research awards - Philips, Mars, Haleon - https://www.wearehuman8.com/news/human8-named-finalist-in-three-global-research-awards-for-human-centric-innovation-with-philips-mars-haleon/