Observed behaviour
AI has become the new tool for food vigilance: more conversational than rating apps, more contextual.
⚠ Tech-assisted food vigilance is stepping up another notch.
Sector expertise, Food & Grocery retail
Most people
want to eat better.
Few do.
That's a problem
of design.
For 10 years, we have been studying real food behaviour, not stated intentions. We know why people don't do what they say, and how to design products, services and experiences that genuinely resolve their everyday tensions.
Signal · 10 years of food fieldwork
15+
social science studies produced in food and retail.
unknowns studies
150+
ethnographic interviews with shoppers across every way of doing the groceries.
unknowns fieldwork programme
8
countries covered by our research: France, Belgium, Spain, Italy, Poland, Romania, Brazil, Argentina.
unknowns international research
Our experience
From shopping trolley to plate,
the food sector
in depth.
From an ethnographic study of how people use Yuka to Intermarché's omnichannel transformation, from European consumers' behavioural tensions to Buffalo Grill's brand platform, we cover the whole food chain, from shopping to the meal.
- Carrefour
- Intermarché
- Buffalo Grill
- Roche Diabetes Care
- INCa
15+
Social science studies on food
150+
Ethnographic interviews with shoppers, across all shopping habits
100+
Interviews with retail staff (store managers, team members, supply-side managers)
1500h
Spent in nearly 30 stores: order picking, delivery rounds, shelf stocking
AI Observatory, Food module
Dedicated “food & grocery shopping” module.
Data updated twice a year: how people turn to AI for shopping, cooking, nutrition and food vigilance.
Insights
What we have learnt
that traditional studies
don't always see.
Our ethnographic research has built a proprietary body of knowledge on real food behaviour.
Our core conviction
Eating is as much a social, identity-driven and symbolic act as a biological one.
People don't manage their diet, they manage their lives. Every purchase decision results from a multidimensional system of constraints: budget, time, cooking skills, each household member's tastes, social norms, moral pressures. Self-reported studies only see the tip of this iceberg.
Our approach: observing what people actually do, in their kitchens, in supermarket aisles, in their daily lives.
Our proprietary modulesMain shopper · store ecosystem · food vigilance · turning points · product trial logic · household dynamics.
Examples of tensions we have modelled
Food behaviour is shaped by deep tensions that no one helps people resolve.
Budget vs. quality
Wanting to eat better vs. not being able to afford it.
The most universal tension. “Eating better” is seen as reserved for those who can afford it, which breeds guilt with no way out.
Individual vs. group
Individual food aspirations vs. eating together.
Every individual preference clashes with commensality, the shared meal, a cardinal value in Catholic cultures.
AI & the plate
AI gives advice
on food.
It doesn't resolve
the real tensions.
AI is quickly settling into the kitchen and the weekly shop. It equips vigilance but resolves none of the underlying tensions, and sometimes creates new ones.
Product tension
AI says “eat less processed food” without solving the household's budget, time or tastes.
⚠ Advice without solutions breeds resignation.
Opportunity
AI can finally tailor advice to the real household.
✦ Personalisation at scale becomes realistic.
Retail stake
The store remains irreplaceable for trying, discovering and the social side of shopping.
✦ The store has an edge AI can't take away.
Case studies
Two engagements.
One common thread:
understanding before
designing.
Carrefour, Europe
Eating better: decoding European consumers' real tensions to uncover 13 repositioning opportunities.
“The problem isn't that people don't want to eat better: it's that “eating better” creates conflicts no one helps them resolve. It's not a willpower problem, it's a design problem.”
Synthesis of 80+ academic and industry sources + CSR data from 4 countries + ideation workshops with country marketing teams. Modelling a system of core tensions. AI-powered concept prototyping.
Outcome
13 prioritised opportunities, prototyped with AI, testable and adaptable across 4 countries. A shared narrative arc for “eating better” positioning at European scale.Intermarché, 4 years
4 years to reinvent the grocery shopping experience: from customer understanding to omnichannel transformation.
“Online grocery isn't a channel: it's a new way of doing the shopping. Satisfaction is won in store, not on screen. The job wasn't to build a website, but a programme that reinvents what it means to be a retailer.”
4 years and 20+ engagements: ethnographic studies, in-store observation, omnichannel strategy, experience design, platform product management, organisational transformation and IT support.
Outcome
RED programme launched (+€1 billion in digital revenue), Digital Department created, e-business platform rolled out.Benchmark study
How Yuka is used in everyday eating habits.
Ever more users, but real behaviour far more nuanced than simply following the recommendations. Our ethnographic study documented why consumers don't automatically change their habits when faced with nutritional information, and what that means for anyone who truly wants to influence food behaviour. We present it on request.
Opportunities
Four engagements we could
take on together.
- 01Strategy & positioningIdentifying your consumers' real tensions and designing a positioning that genuinely helps them.
- 02Experience & retail transformationDesigning a shopping experience that meets real constraints, in store and online.
- 03AI & personalisationDesigning the food AI service that finally resolves the tensions today's tools amplify.
- 04Food transitionSupporting your customers through lasting change: the right moments (turning points) and the right levers (price, point of sale, support).
FAQ
What we're asked
most often.
By starting from what consumers actually do, not from the technology. unknowns designs tools that support concrete uses: AI-powered concept prototyping, advice personalised at household level, agents trained on our behavioural research data. All grounded in a detailed understanding of what people really do with their food, beyond their stated intentions.
AI is becoming the new tool for food vigilance: more conversational and contextual than rating apps, it slips in between the brand and the consumer. But mostly it hands out generic advice (“eat less processed food”) without solving the household's real constraints of budget, time or taste. For a brand, differentiation now lies in the ability to understand and resolve these tensions, ground unknowns has been working for ten years.
A food tension is a conflict consumers live with every day without being able to resolve it on their own: wanting to eat better but not being able to afford it, reconciling everyone's tastes with a shared meal. unknowns has modelled a complete system of these tensions from its ethnographic research. At a time when AI multiplies advice without solutions, this is the asset that makes it possible to design positioning and services that genuinely help people.
unknowns is neither a communications agency nor a market research firm. We combine social sciences, design and generative AI to start from real food behaviour (what people do in their kitchens and supermarket aisles), not from what they say in a survey. This field understanding, built over ten years with consumers and retailers, is what makes our recommendations actionable.