AI-native design and build
Role: Senior Product Designer, Product co-creator
Skills: AI-native product design, Product strategy, UX, Prototyping, Cursor, OpenAI, Front-end experimentation
Context: Independent AI-assisted product development and experimentation
Product focus: Loop Lists, working title. Built in collaboration with Roy Muir.
Output: Working product, staged feature releases, UX flows, product strategy and AI-assisted implementation
Status: In active development
Over the last 12+ months I’ve been working hands-on with AI across product design, prototyping and code, using self-initiated product work to deepen my capability with AI-native tools and workflows.
My main focus has been Loop Lists, a working-title AI-assisted grocery planning app I’m building with a colleague to help busy households, especially families with young children, stay organised, save time, reduce food waste and stretch their grocery budget further.
Using ChatGPT, Cursor, OpenAI and Figma Make, I’ve been moving from idea to prototype to working software faster, while testing product decisions and implementation trade-offs earlier.
Alongside Loop Lists, I’ve explored broader AI-assisted design workflows, including prototype critique, concept refinement, Figma Make prompts, Funding Explorer ideas, structured learning and emerging AI product patterns.
From product thinking to working software
The shift has been from static design outputs to working product experiments. Using AI-native tools has helped me explore ideas faster, test behaviour earlier and make more informed product decisions.
It means I can pressure-test not just how an experience looks, but how it works, where it breaks and what needs to be simplified.
- Frame the product problem and user need
- Explore directions quickly with AI-assisted critique and iteration
- Prototype flows, states and edge cases
- Use Cursor Plan Mode to break features into buildable steps
- Test working behaviour, refine the UX and improve implementation details
- ChatGPT for product thinking, critique, content design and workflow refinement
- Cursor for implementation planning, feature builds, debugging and code iteration
- OpenAI APIs for AI-assisted product behaviour and fallback logic
- Figma Make for rapid UI exploration and prototype direction
- Role-based Cursor skills for PM, designer and developer review
Loop Lists is the main product where I’ve been applying AI-native design and build workflows. It is an AI-assisted grocery planning app for busy households, especially families with young children, focused on helping people stay organised, save time, reduce waste and make more budget-friendly shopping decisions.
- Designed and built working list creation and shopping flows
- Shaped recipe URL parsing from user need through to implementation detail
- Explored smart suggestions based on shopping history and likely repeat needs
- Worked through AI fallback logic, error states and content design
- Explored lightweight planning and budget-friendly grocery concepts
Alongside Loop Lists, I’ve used AI to support design critique, concept refinement and rapid prototype direction. This has included analysing existing prototypes, reviewing them against a brief, refining the product story and creating Figma Make prompts to generate improved design directions.
These experiments have been valuable for learning how AI can act as a design partner: helping challenge assumptions, sharpen flows and move more quickly from critique to an improved product direction. It has also been a useful way to deepen my capability with AI-native design tools and workflows.
The core list was intentionally kept lightweight: quick entry, clear scanning, and thumb-friendly interactions for one-handed use. A subtle completion celebration adds delight at the end of the journey.
To reduce the friction of recipe-to-list planning, I worked on a recipe URL import feature that turns recipe pages into structured shopping list ingredients.
The work included deterministic extraction, AI fallback logic, ingredient normalisation, metric-friendly quantities, blocked-site handling and clearer error messaging when a page cannot be accessed.
I’ve also been exploring how shopping history can support lightweight, helpful suggestions without making the product feel overly automated.
The work includes likely-purchased tracking, purchase-pattern summaries and a “You might also need” experience that can suggest repeat items once enough behavioural signal has built up.
I’ve continued exploring how AI could extend the Funding Explorer concept, helping small businesses search, compare and understand funding options more easily.
- AI-assisted search and filtering
- Structured comparison outputs
- Guided prompts to help users understand options and trade-offs
Early exploration of adding an AI layer to help explain work and provide additional context.
- Structured content prompts
- Context-aware responses
- Lightweight integration approach
- AI-assisted UX for useful, lightweight product experiences
- Moving from product strategy to working software faster
- Using AI to critique, plan, prototype, build and refine features
- Designing around structured data, fallback states and edge cases
- Working more fluently across UX, front-end behaviour and implementation trade-offs