Challenges
What Front + Center Needed to Solve
Problem 1
Automate Cabinet Measurement
Customers needed a way to capture accurate cabinet dimensions without a tape measure or professional help. The core requirement was a model that could read a kitchen photo and return reliable measurements for every opening – by wall, by door, and by drawer.
problem 2
Remove Manual Product Selection
Browsing a product catalogue without knowing exact dimensions leads to wrong purchases and returns. Front + Center needed a recommendation engine tied directly to measurement output so every suggestion matched the actual cabinet configuration.
Problem 3
Guide Users Through the Process
Not every customer knows what information a cabinet order requires. A structured questionnaire was needed to collect style preferences and confirm measurement details before any product logic ran.
Problem 4
Connect Measurement to Checkout in One Flow
The buying journey was broken across too many steps. The app needed to carry data from the scan stage through to cart and order confirmation without asking the user to re-enter anything or switch between tools.
Technical Challenges
Where the Technical Challenges Were
and How We Solved Them
Real kitchens vary in lighting, layout, and cabinet style – and the entire product promise depended on getting the measurements right every time.
01
Detecting Cabinets in Uncontrolled Environments
User-submitted kitchen photos vary widely in lighting, angle, and cabinet style – a model trained on controlled images would fail in real homes. Intuz trained the YOLOv8 object detection model to handle the full range of real-world kitchen conditions, refining it iteratively until detection stayed consistent regardless of how or where the photo was taken.
02
Extracting Accurate Dimensions from a Photo
Automatically measuring cabinet length, width, and height from a single photo required a level of precision that standard image processing does not deliver. The AI/ML model was built and trained specifically for cabinet dimension detection, eliminating the need for manual measuring tapes and ensuring every measurement is accurate enough to drive product selection.
03
Matching Measurements to the Right Products
Cabinet openings vary in size, and a recommendation engine that cannot account for that variation returns mismatched results. Intuz built a recommendation layer that maps detected dimensions to compatible products and filters further by style preferences captured in the questionnaire – so users get suggestions that fit their specific configuration, not a generic catalogue.
04
Keeping the Full Journey in One App
Connecting AI detection, a guided questionnaire, product recommendations, and WooCommerce checkout into a single mobile flow required every step to hand off cleanly to the next. Intuz built the app so that data captured at the scan stage carries through to cart and order confirmation without the user re-entering anything at any point.
05
Real-Time Order Sync Between App and Backend
Order status needed to update accurately after purchase and stay consistent through the full order lifecycle. WooCommerce powers the order management layer, and the sync between the app and backend keeps order data accurate from checkout through to progress tracking and history – all within the same app the user started in.
Have a Complex Product Flow
That Needs AI Built Into It?
Intuz works with businesses to replace manual, error-prone steps with AI systems that are accurate, fast, and built for real users.
Solution
From a Kitchen Photo to a Confirmed Order
Intuz handled the full build – from AI model training and mobile app development to product recommendation logic and order management integration.
The result is a single mobile experience where a customer can scan their kitchen, get accurate measurements, find the right products, and place an order without leaving the app.
AI-Powered Cabinet Dimension Measurement
The app uses a YOLOv8 object detection model to scan a kitchen photo and automatically tag every cabinet opening by wall. It captures width and height for each opening – doors and drawers – with high precision, eliminating the need for a manual measuring tape. Users see every measurement mapped back to the exact opening it came from, wall by wall.
Interactive Guided Questionnaire
After the scan, the app walks users through step-by-step questions to confirm cabinet details and capture style preferences – including front style and finish. Built using React Native, the interface guides users through each input clearly, ensuring the measurement data and style choices are accurate before any product recommendation runs.
Real-Time Cabinet Detection and Personalised Recommendations
Using computer vision powered by YOLOv8, the app analyzes cabinet dimensions and style to surface products that match the specific layout scanned. Recommendations are tied directly to what the model detected – not to manual inputs or catalogue browsing. Users get products that fit their exact space without any additional calculation on their end.
One-Tap Checkout, DIY Resources, and Reassurance Options
The app connects to PayPal for secure payment processing, letting users go from product selection to confirmed order in a few taps – no extra calculations or manual adjustments required. A built-in DIY resource library covers measuring, installing, and painting to support customers through the full refacing process after their order is placed. Reassurance options are available at checkout, giving customers clear coverage based on the value of their order.
Smart Order Management
WooCommerce powers the order management layer behind the app. Users can view recent and past orders, track order status in real time, and see a full breakdown of cabinet details – door quantity, drawer quantity, opening specs, finish, and hinge – all within the same app they used to measure and select products.
Have a Product Idea That Needs AI to Work?
We help product teams identify exactly where AI can remove friction – and then build it.
Tools & technologies
Tools and Technologies We Used
The stack was selected to support real-time AI inference on a mobile device, a guided user flow built for non-technical customers,
and a stable commerce and order layer running behind the scenes.