Introducing EatAI - a customisable AI powered meal planner. Use a ChatGPT-style chat interface to tailor your preferences and receive a personalised meal plan. The meal plan is fully adaptable; for example, you can request adjustments like, “I like that, but can you add more fish for this week?” Additional features include an editable shopping list for your meal plan within a convenient app.

As a new project, I’m designing a new product or brand in four day sprints. EatAI is the first one. Why? Read more.

Problem:

I received instructions from my nutritionist, which are hard to follow. I want to plan a weekly meal plan and head to the supermarket to get what I need.

Solution:

I asked ChatGPT to analyse my nutritionist’s notes and give me a meal plan:

This was pretty cool, but there were a few things I didn’t like. This is where I spotted an opportunity beyond the huge time save ChatGPT gave me.

So I have a great meal plan, I then asked to create a shopping list and headed to the supermarket.

Opportunity

This was fun, and cool, yet the chatbot interface lacked engagement and functionality. I started to consider what this would look like as an interface. The challenge would be to build an app maintaining customisability and a more engaging meal plan and shopping list view. The result - the EatAI Prototype:

Process

I've set a timeframe of four days for the design process. This means I don’t get obsessed by one interesting idea, am forced to release, get feedback and move onto something new. Watch the overview of the process in this video:

Day 1: Problem, research and goals: Following a condensed Design Sprint process on Miro. Miro Board ↗️

Day 2: Sketching: Fun times drawing out ideas on the Remarkable. Uploaded to Miro ↗️ 

Day 3: Prototyping: Designing in Sketch (sidenote: realising that its much quicker to release in Figma but designing in Sketch is faster, time to upskill!). Sketch File ↗️

Day 4: Release: Launching the product and this blog post. Try the prototype ↗️ 

This process allows me to take an idea and rapidly launch something to the world. By having a process, I can review it at the end and retrospect on how to make it better. By launching something in four days, I can quickly review and launch better the next time.

Results

If you’ve checked out the prototype, you’ll see there’s a link at the end to leave a review on a Google form. I have no idea if anyone will fill this out, nor is it a good way to do research! (side note: I recommend person to person interviews, here’s a write up of a process we followed at my last design studio, Deep Work and a client example). Anyway, the results are live as they get filled in: Google Form Results ↗️

Notes on the process

If you’ve read this far, that’s pretty cool! You’ll also note there’s lots of things that aren’t very high quality. (Apart from the UI design in the application, I am trying to practicing here after all). My objective is to optimise the process for launching in the shortest time possible to get feedback and move on to something new. This means things like not doing high quality user interviews or exploring other design directions.

Try the prototype

Please give it a go and fill the feedback in at the end:

What’s next?

Follow my public notion to see what I’m doing and progress report live.

Keep subscribed on Substack for the next launch.