Akaani
Designing the AI-powered Nigerian food tracking experience from 0 to 1.
- Role
- Lead Product Designer
- Timeline
- 2024 - Present
- Platform
- iOS & Android
- Year
- 2024
The Problem
Nigerian food is deeply cultural, but no macro tracking app spoke to it. Jollof rice, egusi soup, and suya do not have barcodes. Nutritional databases built for Western diets leave Nigerian users guessing, entering "white rice" when they mean ofada, picking "chicken stew" when they mean ayamase. The result is data so inaccurate it stops being useful.
Akaani was built to fix that. An AI-powered food tracker built specifically for Nigerian and African meals, with a conversational assistant named Lu at its core. The mission: help Africans eat better, on their own terms, without translating their culture to fit a tool that was never designed for them.
My job was to design the entire mobile product experience, from the moment a new user lands on the app, through the free tier, into premium onboarding, and across the core tracking and planning features.
My Role
I came in as the core product designer, working directly with the CPO and CTO. I owned the end-to-end design: user research, information architecture, wireframes, the full design system, and every screen from freemium onboarding to the premium paywall. I also wrote the UX copy for Lu, the AI assistant, and designed the recipe shop and marketing website.
Team
Lead Designer (me), CPO: Susan Olapade, CTO: Richard Igbiriki, 2 engineers
Research & Discovery
Before touching Figma I needed to understand how health-conscious Nigerians actually think about food tracking, what they try, why they quit, and what they wish existed.
Users quit after 3 days
Every tracker they tried failed within a week. The common reason: spending 10 minutes searching for "egusi soup" only to find a result with clearly wrong macros, or nothing at all. The friction of a bad database broke the habit before it started.
Portion size is the real problem
Nigerian food is not served in neat 100g portions. A plate of jollof rice at a party is different from a home plate. Users knew they were eating well, they just had no way to quantify it in a way that felt honest.
They trust conversation more than forms
When asked how they would prefer to log a meal, almost every user described talking to someone, "I would just say what I ate." This was the core signal that shaped Lu's role in the product.
Search makes or breaks the first log
The fastest path to value is finding your dish in seconds. People name food in many ways, so search had to understand local names, spellings, and short forms: "jollof", "party rice", and "concoction rice" all needed to resolve to the right meal. Get search right and the habit sticks.
Premium value had to be felt immediately
Free users who did not hit a paywall within the first session rarely converted later. The freemium model needed to show value fast and create a moment of genuine desire, not frustration, at the gate.
Key Decisions
The Question
Should macros be blurred or hidden entirely for free users?
The Decision
Blurred, always visible but unreadable.
Hiding macros removes the desire. Blurring creates it. Free users can see the shape of the data, they know the number exists, they just cannot read it. That tension drives upgrades without feeling punitive.
Flow Diagram
The Question
Where should Lu live in the navigation?
The Decision
Dedicated tab, center position in the tab bar.
Lu is not a support feature, she is the product. Putting her in a corner or floating button buried her. Center tab placement signals that conversation is the primary interaction model, not a secondary option.
The Question
How do we handle the annual vs. monthly pricing display?
The Decision
Annual plan shown first, highlighted, monthly shown as the alternative.
Showing monthly first anchors users to a higher price, making annual feel like a downgrade. Showing annual first reframes the monthly option as the more expensive choice. We always lead with annual at $4.17/month, with 'Save 40%' visible before any other number.
The Question
Should streaks be free or premium?
The Decision
Streaks are free, forever.
Streaks are a habit-formation tool, not a value-add. Gating them creates resentment. Keeping them free drives daily engagement across all users, which in turn creates more upgrade moments.
Final Design
The final product spans freemium onboarding, the core tracking experience, Lu's chat interface, premium features, and the paywall system. Here is the full design.
All Screens
Key Flows
First session, free user
Lu chat flow
Outcomes
2,000+
signups before public launch
Organic, no paid acquisition
200+
Nigerian dishes catalogued
With accurate macro data
7-day
free trial with CC-required flow
No charge until day 8
What I Learned
The biggest thing this project taught me is that gating features is a design problem, not a business decision handed down from above. How you gate something (the moment, the copy, the visual treatment) determines whether a user feels respected or manipulated. Getting the blur treatment right took four iterations. The paywall copy took even more. If I were starting again, I would run user testing on the paywall moment earlier. We made assumptions about what would feel fair that we could have validated in week two instead of month three.
Next Project
SaaS Product