Todelli AI Advisor – Nutrition Intelligence + Commerce Integration

End-to-end service and product design for Todelli’s evolution into a contextual nutrition AI experience directly connected to commerce.

Approach

  • Defined core user scenarios, the end-to-end journeys, and shaped how the AI advisor progressively comes to life within the existing platform and later as a standalone agent. I conducted qualitative research (1:1 interviews), validated behavioural insights, and translated them into a personas and user scenarios, plus an initial concept prototype that was further tested with the target audience. The result was clear user driven feedback for a scalable personalised nutrition and shopping solution.

  • Macro Trends Supporting the Proposition

    Shift toward organic, clean-label, plant-based diets

    Studies across 2020–2024 show:

    • Significant rise in positive sentiment and consumption of organic foods, expanding from niche to mainstream.

    • Post-COVID populations demonstrate higher interest in food safety, nutritional value, and ethical sourcing.

    • Demand for clean-label, plant-based alternatives, and functional foods continues to grow year-on-year.

    Growing desire for transparency & contextual nutrition guidance

    Consumers increasingly want:

    • Ingredient clarity

    • Sustainability and ethical production information

    • Personalisation based on goals (energy, gut health, hormones, longevity)

    Commerce gap

    No major nutrition platform currently:

    • Provides contextual nutritional advice at the moment of product discovery, AND

    • Links that advice to shoppable products / personalised baskets / recommendations.

    This unmet need creates a commercial opport

  • Information Overload

    Organic, plant-based and functional foods come with complex claims. Users struggle to understand:

    • What’s good for their body

    • How to evaluate quality and processing

    • What aligns with their goals (gut, hormones, energy, inflammation, weight)

    Fragmented Tools

    Current solutions are either:

    • Diet apps (calorie trackers, food journaling)

    • One-dimensional wellness tools (gut or hormone focus)

    • E-commerce stores lacking nutrition guidance

    • Nutritionists who are expensive or inaccessible

    No solution connects:
    Real-time nutritional context with personal advice and purchasing.

    Loss of Consumer Trust

    Packaging claims are confusing; many wellness brands overpromise.
    Users want a trusted, evidence-based advisor, not another marketing engine.

  • A conversational AI agent that delivers:

    Contextual Product Intelligence

    When browsing food items, users receive:

    • Ingredient clarity

    • Health impact explanations

    • Sustainability and ethical insights

    • Nutritional comparisons

    • “What this means for you” advice

    Personalised Guidance

    The agent can answer:

    • “Is this good for my gut?”

    • “What can help balance hormones?”

    • “Compare this kombucha to that probiotic drink.”

    • “Create a weekly plant-based, iron-rich plan.”

    Commerce Integration

    Seamlessly connects recommendations to:

    • Todelli marketplace products

    • Personalised bundles

    • Smart substitutions

    • Automated shopping lists

    Phased Product Evolution

    1. Phase 1 — Embedded Intelligence
      Integrated into existing Todelli app/website, supporting browsing.

    2. Phase 2 — Proactive Scenarios + User Journeys
      Advisor initiates help based on user context (e.g., scanning a product, searching for recipes).

    3. Phase 3 — Standalone AI Agent
      Independent, conversational experience: talk, write, ask anything nutrition-related.

  • Objectives

    • Validate user desirability

    • Validate willingness to pay

    • Identify top use cases

    • Proven ability to drive higher conversion

    • Measure trust and perceived value of AI guidance

    Testing Components

    1. Interactive Prototype Testing
      With 20 nutrition-passionistas (completed)

    2. Live Experiment Inside Todelli Platform
      Limited AI advisor access to 10–20% of users.

    3. A/B Tests
      AI vs no AI on:

      • Product understanding

      • Conversion to purchase

      • Time spent

      • Basket size

    4. Qualitative Interviews
      Understanding emotional trust factors & psychological value. (Complete

    Success Metrics (KPIs)

    • +15% product discovery-to-purchase conversion

    • +25% engagement time

    • 40%+ willingness to pay for subscription features

    • 60%+ trust score in AI explanations

    • Top 3 high-value user scenarios identified

    Outcome of Phase 1:
    Green-light for robust development & investor pitch with validated metrics.

Target Audience

Primary: “Nutrition-Passionistas”

  • Age 25–54

  • Strong interest in clean eating, plant-based lifestyles, organic produce, sustainability

  • High digital adoption

  • Willing to pay for quality products

  • Seek personalisation, learning, and convenience

Secondary: “Wellness-Motivated Mainstream Consumers”

  • Occasional organic buyers

  • Curious about functional foods

  • Need help navigating claims and brands

Personas & User scenarios were developed in remote co-creative workshops along with the rest of the team post research.

Visual concepts and prototyping

I used Figma and UX Co-pilot to bring the concept to life and create a quick prototype (for core user scenarios) that was tested with our target audience. Based on the insights we amended the following:

Shift from open-ended chat to:

  • Quick-select buttons

  • Pre-filled prompts

Why

Users who tested the early chat prototypes said:

“I’m not sure I would always know what to ask.”
“Sometimes this can feel like work.”

AI responses became

  • More structured

  • Shorter

  • With clear headings (“For your energy goals…”)

  • Bullet-based rather than paragraph-based

Why

During testing, users often skimmed the replies and miss key info.
Clarity directly impacted trust and perceived intelligence.

Initial MVP Iteration

As a result of the research, and to further validate our approach with the marketing and analytics team, I designed an improved product page with nutritional AI embedded into the product information. The page comes to life through an integrated chat agent available to users. It includes nutrition facts that are often not highlighted on traditional labels—such as fibre and protein—along with their nutritional significance (for example, whether levels are low or high). The page also provides health benefits of the product, usage guidance, and explains how it can serve as an alternative to common ultra-processed foods (UPFs) – aligning with our catalogue of championed products that are good for both personal health and the planet.

Next
Next

Blackghost