Challenges

Most Users Already Know What to Cook

Receipt scanning isn’t useful—users already know what they’re cooking, so they don’t need AI to suggest recipes from their purchases.

AI Needs More Than One Image

Because food is often hidden behind other items, one photo of the fridge or pantry isn’t enough for AI to create an accurate ingredient list.

Making Meal Prep Seamless for Users

After speaking with target users, the design was updated to reflect their needs, including shopping lists, ingredient images, and recipe customization options. Rather than making weekly meal prep a separate pop-up, we decided it should appear directly whenever users access the camera in the app.

Solution

Streamlining Features on the Camera Screen

The final solution combines all features on the camera screen for a more integrated and intuitive experience.

Building a Prototype with Figma Make

I used Figma Make to create a clickable prototype, iterating multiple times to fix flows and refine screens.

This prototype helps developers and stakeholders understand the intended experience and explore it themselves. Even though it’s not a built product, it effectively communicates the full vision.

Helping Users Get Started
Since users might not know where to start the first time, we included a tooltip-based first-time UX tutorial to guide them.

User Testing & MVP

After identifying the UX challenges, we consolidated the flow into a single, flexible interface and completed an MVP. Instead of testing each screen individually, we created a one-pager showcasing the four main screens, explaining the benefits of each, and allowing users to understand the overall experience quickly.

We used DScout to recruit qualified participants, targeting:

• Families who cook regularly

• Users with children (primary meal planners)

• Participants who trust and feel positive about AI

We disqualified participants who didn’t fit the target audience or were skeptical of AI, to ensure feedback was relevant.

Participants answered a mix of basic and in-depth questions before seeing the one-pager, including:

• Do you meal prep?

• Who does most of the cooking?

• What tools and processes do you use for cooking?

• What do you like and wish for in a meal planning app?

Team Goal & One-Pager Approach

User Testing Approach

Our goal was to validate the core feature without distractions from design details. Full prototypes often make users focus on fonts or flow, derailing feedback. The one-pager kept attention on whether the feature is useful, valuable, and helpful for their family, giving us clear product validation.

Targeting attributes

Figma mvp design

1. Age: 25-55

2. Country: United States

3. Household income: $75k+

4. Location type: Urban or Suburban

5. Employment: NOT retired

6. Parental status: Parent/guardian to children under 18 living at home

7. Household composition: With partner/spouse + child

User Feedback

Understanding Users Through Their Words

"So if somehow I could take a picture of my cupboard, and it would document everything I have or even if I had to manually input it, and then I could it would remove it as I used a recipe."

Suggesting meals based on what’s in my fridge/pantry

52%

The top preferred use cases for AI support with meal planning include

(before showing users the one pager)

Suggesting quick or low-effort meals

"If I could just save time and reduce our stress and just help everyone just be happy and eat well, it would be really good for our household."

Finding new recipes that fit our needs/preferences

"So planning meals that, satisfy everyone can be a little hard sometimes. So I would want the AI to suggest meals that balance all of our preferences."

45%

41%

"The fact that it can generate things I already have in my fridge is very appealing, so I don't waste any food."

Top 3 preferred features that mentioned by users.

(after users reviewed the one pager)

Use what (ingredients) you have

"So if somehow I could take a picture of my cupboard, and it would document everything I have or even if I had to manually input it, and then I could it would remove it as I used a recipe."

"I really like the 'personalize your recipe search.' Seeing something like high protein really catches my attention. It makes me think that the AI planner could give me helpful recipes to accomplish my health goals."

One-shot inspiration

"What stands out most to me as I see this for the first time is the ability to take a picture of what’s in your fridge and have AI help create a meal plan."

Survey Respondent

Customize your recipe search to fit specific budgets, nutrition goals, or dietary needs.

Generate meal ideas based on what’s already in your fridge or pantry.

Personalized recipes

Take any idea or image and instantly turn it into a recipe or meal plan.

likely to try the app and find it appealing.

97 %

90 %

are willing to pay for this app.

What We Learned From User Interviews

79 %

feel this app idea is very unique. 69% have never heard of an app like this before.

$6-10/month is the most popular price point (41% of respondents).

Users believe the app can save time, reduce meal-planning stress, and offer helpful personalized recipes based on what they already have at home.

Users highlighted the app’s unique ability to generate personalized recipes from pantry items or photos, something they haven’t seen in other meal planners. This AI-driven “start with what you have” approach stood out as the key differentiator.

How Feedback Informed the Next Iteration

The insights from our users serve as a lighthouse, guiding us toward a better UX. Reading and watching their feedback was both exciting and inspiring. By incorporating these insights, we can continuously refine our design and create products truly centered on users’ needs.

DESIGN OPPORTUNITIES

• Allow users to thumbs up or down generated recipes to improve personalization.

• Support extensive customization, reflecting not just preferences but also dislikes.

• Learn from meal history to understand what users and their families like and dislike.

• Generate shopping lists and suggest ways for family members to contribute ingredients.

• Share recipes easily with spouses and other family members.

Final (Updated) Designs

These screens highlight key updates based on user interview insights, focusing on what matters most to them.

AI-Powered Recipe Discovery

• Finds recipes that match ingredients users already have at home.

• Recipes come from reputable cooking sites with high reviews, ensuring tasty results.

• Users can filter recipes by what they want or need to avoid, a feature users valued highly in research.

Recipe Details & Collaboration

• Shows ingredients, instructions, and nutrition info

• Highlights what percentage of ingredients users already have.

• Users can check off ingredients, see who’s bringing what, or buy missing items via Instacart.

• Recipes can be shared with others cooking together.

Recipe Voting & Family Preferences

• Users can thumbs up or down recipes, saving likes to their meal history.

• Helps AI learn family preferences and prioritize popular recipes.

• Supports meal planning by showing which recipes are most liked by the family.