SMME E-commerce Platform
Designed an AI-driven discovery experience for Strove that turns plain-language buyer requests, typed on the web or sent over WhatsApp, into precise matches with the right small business, while giving micro and small business owners a guided, low-friction way to get found and get paid.
To help South African buyers discover the right small business for the job in seconds, and give SMMEs a credible, low-friction storefront that surfaces automatically when they are the best match.
Business Problem
South African buyers do not search the way listing directories expect them to. They type "dog groomer sandton" or describe a symptom, not a category and a filter list. Meanwhile SMMEs lack the time, technical skill, or capital to build a discoverable storefront, leaving them dependent on informal channels and social media with no reliable way to be found by the right buyer at the right moment.
Approach
Designed the buyer-facing discovery experience around a resolver-and-ranker model: natural-language queries are interpreted against a structured category and facet taxonomy, then matched deterministically against merchant offers, reviews, proximity, and price, so results stay explainable and consistent across the website and WhatsApp. Paired this with contextual research into SMME owners' digital literacy and a guided, mobile-first onboarding flow that lets owners describe their business in their own words and have facets pre-filled for confirmation, rather than filling out a long structured form.
Buyer Discovery Research
Studied how South African buyers actually search for local services and products: overwhelmingly lazy, symptom-led queries rather than structured category browsing. This shaped a design principle: the experience must be excellent at "dog groomer sandton" and merely capable at a fully-specified request, since that is the real distribution of buyer intent.
AI-Driven Search & Facet Design
Designed the buyer-facing interaction model around a two-pass search: an instant, lightweight match on every keystroke, and a richer clarification pass only when it can genuinely change the result, never more than two clarifying questions on web. Designed facet chips, provisional results shown alongside any question, and an "unsure" option that always keeps a buyer moving rather than blocking them.
SMME Owner Research & Onboarding Design
Conducted contextual interviews and shop visits with SMME owners across retail, fashion, and craft sectors to understand digital literacy levels and existing workarounds. Designed a guided, mobile-first onboarding flow where owners describe their offer in plain language and confirm AI-suggested facets and pricing with a tap, rather than manual data entry, cutting listing effort dramatically for first-time sellers.
Trust, Ranking & Transparent Matching
Designed the buyer-facing logic and UI for how match quality is communicated: review-derived quality signals, proximity, price fit, and availability, so a buyer always understands why a business was recommended. Where no perfect match exists, designed an honest near-miss pattern that explains the gap rather than returning an empty result, and worked with the product team to embed trust signals like verified seller badges and transparent fee breakdowns.
