KikyCodes/M-Pesa-Smart

M-Pesa Smart Budgeting Assistant Designing a conversational budgeting layer for Kenya's most-used mobile money platform — bringing financial clarity to 32 million daily users. # M-Pesa Smart Budgeting Assistant — UX Case Study A full UX case study exploring the design of a native AI-powered budgeting feature for M-Pesa, Kenya's dominant mobile money platform serving 32M+ daily users. --- ## Overview M-Pesa processes billions of shillings daily, yet offers users no meaningful view of where their money goes. This case study designs a **Smart Budgeting Assistant** — a contextual AI layer built directly into the M-Pesa app — that uses real transaction data to automatically categorise spending, surface insights, and offer proactive financial guidance in Swahili and English. --- ## What's Inside | Section | Description | |---|---| | **Problem Statement** | Market context, the financial visibility gap, and key stats | | **User Personas** | 4 research-driven personas across Nairobi, Kisumu & Nakuru | | **Pain Points** | 8 validated user frustrations mapped to behaviour patterns | | **User Journey** | Full salary-cycle journey map for primary persona Wanjiku | | **Solution** | The AI budgeting layer concept with phone UI mockup | | **Feature Breakdown** | 6 features, each traced to a specific research pain point | | **Success Metrics** | Post-launch KPIs targeting behaviour change, not just engagement | --- ## Personas - **Amara Njeri** — Freelance designer, variable income, high tech comfort - **Kevin Otieno** — Matatu driver, daily cash, family remittances, USSD-first - **Wanjiku Muthoni** — Salaried teacher, savings goals, impulse spending struggles - **James Kimani** — Hardware shop owner, business/personal funds blurred --- ## Key Features Designed 1. **Auto-Categorisation Engine** — Tags every transaction using till numbers, paybill codes, and recipient patterns. No user input required. 2. **Payday Budget Planner** — Salary-triggered flow that pre-fills a budget from last month's spending in under 60 seconds. 3. **Conversational AI Coach** — Natural language Q&A ("Nimetumia pesa ngapi kwa chakula?") powered by live transaction d …