Finding a properly fitting shoe has traditionally relied on rough sizing standards that vary between brands and often fail to account for individual differences in foot shape, arch height, and gait pattern. Artificial intelligence is now being applied across the footwear industry to close this gap, from in-store fitting kiosks to fully custom, algorithmically designed manufacturing.
Computer Vision Foot Scanning
A growing number of retailers now offer AI-powered foot scanning using nothing more than a smartphone camera. Computer vision algorithms analyze dozens of measurement points from a few photos, capturing length, width, arch height, and even subtle asymmetries between the left and right foot, data that previously required specialized in-store equipment to obtain accurately.
Predictive Fit Algorithms
Beyond simple measurement, machine learning models are increasingly used to predict how a specific foot shape will interact with a specific shoe model’s internal geometry, accounting for material stretch, seam placement, and known pressure points. Some platforms can now recommend not just a size, but a specific width and model most likely to fit well based on patterns learned from thousands of previous fittings and returns data.
Gait-Informed Design
AI systems are also being used to analyze video or sensor-based gait data and translate it into specific footwear recommendations, identifying overpronation, supination, or asymmetric loading patterns and matching them to shoes with corresponding stability or cushioning features. This moves fitting technology beyond static foot shape into dynamic movement analysis, arguably a more clinically meaningful measure for injury prevention.
Generative Design for Custom Midsoles
On the manufacturing side, generative design algorithms can now create midsole lattice structures optimized for an individual’s body weight, activity type, and pressure map, computationally generating support structures that would be impractical for a human designer to engineer manually. These designs are then produced using 3D printing, creating a genuinely individualized product rather than a mass-produced shoe with a swapped insole.
Reducing Returns and Waste
From an industry perspective, AI fitting technology addresses a significant practical problem: footwear has historically had among the highest online return rates of any retail category, driven largely by fit issues. Better predictive fitting reduces both the inconvenience to consumers and the substantial environmental waste associated with shipping, returning, and often discarding ill-fitting shoes.
Health Applications Beyond Retail
Clinically, AI-assisted fitting is being explored for patients with diabetes, arthritis, or post-surgical foot changes, where proper fit is not just about comfort but directly tied to ulcer and injury prevention. Some healthcare systems are piloting AI foot-scanning kiosks in clinics to streamline the prescription process for therapeutic and diabetic footwear, which traditionally involved lengthy, manual fitting appointments.
Current Limitations
Despite rapid progress, AI fitting technology still has real limitations. Smartphone-based scans can be affected by lighting, camera angle, and user error, and predictive algorithms are only as good as the fit and return data they’re trained on, which can carry biases from limited brand or demographic representation. Most experts still recommend an in-person fitting for people with significant foot conditions rather than relying on AI recommendations alone.
Where This Is Headed
As sensors, cameras, and machine learning models continue to improve, the gap between digital fitting predictions and real-world comfort is narrowing quickly. The long-term trajectory points toward a footwear industry where standardized sizing becomes a fallback option rather than the default, replaced by individually optimized shoes designed around each person’s unique foot shape and movement pattern.