Home AI Thought Prototypes Contact

AI Thought Prototypes — RIP™

Commercial problems,
held long enough
to build something real.

Every one of these started as a question I could not stop thinking about. A category that confuses rather than converts. A science nobody makes legible. A behaviour every brand tries to change but nobody properly understands. AI removed the distance between the question and a testable answer — and I have embraced it with wide arms.


These are working hypotheses — live, interactive, built on real commercial frustrations. Play with them. If something resonates, I would be happy to show you what it takes to build it properly — let’s talk.

My background is brand strategy and marketing, not technology. My strength lies in finding commercially viable ideas and solutions in most complex and ambiguous situations. For years, the distance between having an idea and being able to test it required agency and R&D briefs, big budget, and long development timelines. With AI, that gap no longer exists in the same way — and I have embraced it with wide arm.

I call this RIP™ — Rapid Idea Prototyping. It’s an approach where I take one lived experience, that has a clear consumer / brand challenge, a white space opportunity, but not fully solved yet. I use one or two proxy categories to build the proof of concept, but the application is relevant for Consumer goods, Retail, Pharmacy, FMCG, E-commerce, wellbeing, media strategy.

Here I share few of those now and I have been building this stack continuously.

Four experiments — live and interactive

Consumer Decision

Find Your Fit

The Recommendation Engine

74% of consumers walk away from a purchase because they are overwhelmed by choice, and 41% say it takes more effort to decide than it did three years ago (Accenture Consumer Pulse Survey, 2024 — 19,000 respondents across 12 countries). The problem is — the explosion of choice, poor product navigation support from manufacturers, and the near-total absence of plain-English guidance at the point of decision. The consumer who picks the wrong product first takes years to win back, if ever.

Three behavioural questions — motivation, context, preference. One named product suggestion with a plain-English rationale. The proof of concept is built across two of the most commonly confusing categories — wine and tobacco alternatives — with outputs calibrated by market, starting in the UK. The logic is portable to any category where complexity erodes conversion at the moment of decision.

Where this applies

  • Consumer goods
  • Retail
  • Pharmacy
  • FMCG
  • E-commerce
Open experiment
Hyper-Personalised Wellbeing

Pickform

Botanical Performance Simulator

The global dietary supplements market is projected to reach $402 billion by 2034 (Global Supplement Industry Trends, 2025), with the brain health segment alone expected to more than double to $7 billion in Europe by 2033 (Market Data Forecast, 2026). Over 63% of consumers believe cognitive health is central to their overall wellbeing (Favoured, 2025) — yet the category remains decipherable only by serious biohackers. The label tells you what is in it. It almost never tells you why, in what combination, or whether it is right for your body. And almost nobody accounts for the fact that compounds metabolise differently by age, sex, and ethnicity. Consumers are asked to trust formulas they have no easy way to evaluate.

Select an optimisation goal — Deep Focus, Stress Recovery, Morning Activation, Sleep Onset, Social Ease, or Physical Recovery — and dial four performance parameters, then input your biological profile. The tool uses published research on how compounds interact with different biological profiles to suggest a formula — active ingredients, approximate ratios, preparation notes, and relevant cautions — as a starting point for personal experimentation. A 7-day tracking module lets you log wearable indicators — such as Oura Ring readiness score, HRV, sleep score, and resting heart rate — alongside your own daily notes. The goal is to help you notice patterns, not to establish clinical outcomes. There is no proven 1:1 relationship between these supplements and any specific result. This is a thinking tool, not a prescription.

Where this applies

  • Nutraceuticals
  • Sports performance
  • Preventive health
  • Adaptogens & nootropics
Open experiment
Product Innovation

NotSmoke

5,000 Years of Plant Intelligence

79% of people who attempt to stop smoking relapse within six months (PubMed, 2011 — replicated in multiple subsequent studies). The leading reasons are not nicotine craving alone — social triggers and behavioural discomfort account for over 60% of relapse drivers (Tobacco Induced Diseases, 2023). Most cessation products replace the chemical. None replace the pause, the hand-to-mouth action, or the sensory moment that signals the shift. That is a design failure, not a willpower failure. For thousands of years, societies have used botanicals to manage exactly these moments — from Zen morning incense to pre-dawn tea ceremonies to Indigenous sunrise rituals. The science of why those plants worked is not alternative medicine. It is documented neurochemistry. This tool explores that 5,000-year body of knowledge to find possible alternatives — not just for tobacco, but for any product where the dependency is as much behavioural as chemical: alcohol, sugar, processed food.

Select the product you want to replace, a ritual moment, and a craving profile. The tool maps the reward loop — trigger, ritual, dissolution, neurochemical shift, loop closure — and suggests a plant-molecule protocol intended to address both the chemistry and the behaviour. These are directions for exploration, not clinical substitutes. Every suggested botanical is grounded in a historical precedent timeline — from 800 CE Japan through 1000 CE China, the Americas in the 1500s, and into modern Europe — showing centuries of documented human use, not brand claims.

Where this applies

  • Harm reduction
  • Tobacco
  • Alcohol
  • Processed food
  • Behavioural health
Open experiment
Consumer Obsession

SeeThrough

Three-Dimensional Instant Consumer Intelligence

Nearly 60% of digital FMCG ad spend generates a negative or uncertain return (WARC). Globally, brands wasted an estimated $73 billion on digitally unsuitable communications in a single year (Campaign Live / Campaign Asia, 2023). The inputs that would reduce this risk already exist in most large companies — consumer tracking data, social listening, peer-reviewed behavioural science — but they are never in one place when the agency and brand team are actually creating the content. So teams build on the memory of previous campaigns, rely on boardroom opinions shaped by individual taste, and then commission expensive research to find out, too late, whether the communication works. This is the opposite of consumer obsession. And it is standard practice.

This tool brings those three data points together at the point of creation — not after. Define your target consumer in category terminology. The tool builds a driver and barrier map from three simultaneous signals: your brand’s ongoing consumer tracking data; live social signals including Reddit, reviews, and ratings; and the latest peer-reviewed science. Then paste in your headline, claim, script or visual asset. It scores it on Switcher Resonance (0–100), surfaces verbatim consumer quotes from live communities, and indicates where the communication may be missing the intended audience — before the budget is committed. It does not replace pre-launch comms testing. It gives you a stronger, more grounded brief to take into research.

Where this applies

  • FMCG
  • Regulated categories
  • Campaign planning
  • Media strategy
  • Consumer insights
Open experiment