Case studies

The Problem

SACCOs across Uganda manage millions in member loans using paper ledgers and manual tracking. This creates audit gaps, loan officer errors, and zero real-time visibility. When electricity or internet fails, operations stall entirely.

What We Built

Desktop-first loan management that works fully offline and syncs to the cloud automatically. Covers the complete SACCO lending lifecycle: loan origination, disbursement, repayment scheduling, and financial auditing.

How We Solved It

Offline-first architecture using PowerSync for bidirectional sync between local SQLite and Supabase. Sensitive data encrypted on-device via Electron safeStorage API. Financial audit module computes Portfolio At Risk (PAR), classifies loans by risk level, auto-detects red flags, and generates audit snapshots for regulatory reporting.

Key Outcomes

• Full offline operation with automatic background sync
• PAR computation replaces manual risk assessment
• Red flag auto-detection reduces audit preparation time
• Encrypted local storage protects member financial data
• Role-based access ensures loan officers only see what they need

The Problem

Most Ugandan schools manage academic records in physical files. Report cards are typed manually each term. Parents receive zero communication between meetings. Headteachers make decisions without data summaries. Generic software does not understand the Uganda CBC curriculum or the reality of mixed connectivity.

What We Built

A school management platform that digitizes the full academic cycle — admissions, attendance, assessment, report generation, and parent communication — built specifically for Ugandan school structures.

How We Solved It

Dynamic report card engine pulls subjects, scores, teacher remarks, attendance, and co-curricular entries from Supabase and renders print-ready HTML templates. A WhatsApp AI assistant powered by DeepSeek and Green API gives each parent access scoped strictly to their own children. Headteachers receive a daily AI-generated summary of attendance, performance flags, and outstanding items.

Key Outcomes

• Dynamic report card generation eliminates manual typing each term
• WhatsApp assistant gives parents real-time access to academic status
• Daily AI summary keeps headteachers informed without manual reports
• Student ID cards, meal cards, and admission letters from the same data
• CBC-aligned assessment module reflects the Uganda curriculum structure

The Problem

Small retail shops in Uganda track sales in exercise books. Stock levels are unknown until shelves are empty. End-of-day cash reconciliation is guesswork. Existing POS solutions are too expensive, designed for foreign retail, or require constant connectivity.

What We Built

A lightweight, fast point-of-sale system built to run on low-spec hardware, work offline, print thermal receipts, and give shop owners a clear view of their business without requiring technical knowledge.

How We Solved It

Designed around the cashier workflow — fast product search, barcode scanning, one-tap sale completion. Offline capability is the foundation, not a feature. UGX pricing is native throughout. Mobile money payments via MTN MoMo and Airtel Money are tracked alongside cash. Role-based access keeps Admin, Manager, and Cashier views separate.

Key Outcomes

• Runs fully offline with automatic sync when online
• Daily, weekly, monthly sales reports replace end-of-day guesswork
• Thermal receipt printing works without internet
• Mobile money tracking closes the gap between digital and cash reconciliation
• UGX-native formatting removes currency confusion

The Problem

Transport and logistics companies in East Africa manage vehicle fleets with manual logbooks, phone calls to drivers, and paper fuel receipts. Route deviations go undetected. Maintenance is reactive. Clients receive no real-time shipment updates. Fuel fraud eats into margins with no audit trail.

What We Built

A GPS-enabled fleet management platform that tracks vehicle locations in real time, logs trips automatically, monitors fuel consumption, and sends proactive maintenance alerts — all accessible from a single dashboard.

How We Solved It

IoT GPS units stream telemetry to the backend at configurable intervals. The map layer displays live vehicle positions, geofence breach alerts, and historical trip playback. Fuel tracking compares pump readings against distance travelled to flag anomalies. Maintenance schedules are auto-generated from engine hours and odometer readings. Clients get a branded tracking link per shipment.

Key Outcomes

• Real-time vehicle tracking replaces phone-based driver check-ins
• Geofence alerts notify managers of route deviations immediately
• Fuel anomaly detection reduces fraud by surfacing consumption outliers
• Automated maintenance scheduling extends vehicle lifespan
• Client-facing tracking links improve transparency and reduce support calls

The Problem

Public and private health facilities in Uganda manage pharmaceutical stock using paper bin cards and Excel sheets. Expiry dates are checked manually. Stockouts are discovered only when a prescription cannot be filled. The last-mile supply chain from National Medical Stores to remote clinics has zero visibility.

What We Built

A medical inventory management system that tracks stock levels, expiry dates, and consumption patterns across multiple facilities — with offline capability for remote health centres.

How We Solved It

Each facility manages its own inventory with barcode scanning for receiving and dispensing. Expiry dashboards surface soon-to-expire batches with automatic alerts. Consumption analytics identify usage patterns to prevent both stockouts and over-ordering. The system works offline in rural clinics and syncs when connectivity is available. Central administrators see aggregate stock positions across all facilities.

Key Outcomes

• Real-time stock visibility replaces paper bin card counting
• Expiry alerts reduce medicine wastage from expired stock
• Consumption analytics improve ordering accuracy and reduce stockouts
• Offline sync keeps rural clinics operational without internet
• Central dashboard gives health administrators aggregate supply chain visibility

The Problem

Uganda and East Africa face recurring climate crises — floods, droughts, landslides, crop disease outbreaks — yet response coordination is fragmented. NGOs and government agencies operate from separate data streams and delayed field reports.

What We Built

A command-center-grade monitoring platform that consolidates real-time climate events, natural disaster data, and crop disease outbreaks onto a single interactive map, enriched with AI-generated situation reports.

How We Solved It

Uses PostGIS for geospatial queries and MapLibre for the interactive map layer with per-disaster-type glow markers. Climate Watch dashboard provides a dark command-center aesthetic. AI enrichment via DeepSeek adds contextual analysis. REST API enables NGO and government partners to consume data programmatically.

Key Outcomes

• Unified situational awareness across climate event types
• Early warning alerts for crop disease outbreaks before they spread
• AI-enriched reports reduce analyst time per incident
• REST API enables integration with existing government systems
• Partnership-ready data infrastructure for NGO and donor reporting

The Problem

Commercial farms in East Africa manage multiple fields, crop types, irrigation systems, and drone fleets with no single operational view. Soil conditions are checked manually. Pest outbreaks are discovered late. Irrigation failures are noticed only when crops are stressed.

What We Built

An end-to-end agricultural management platform giving farm operators a real-time dashboard of field health, soil chemistry, drone status, and AI-driven action recommendations — organized by field sector.

How We Solved It

Each farm is divided into named sectors with crop type, health status, and active alerts. Live telemetry — soil moisture, temperature, humidity, wind speed, rainfall, solar radiation — feeds the dashboard in real time. Soil chemistry panel tracks NPK levels with deficiency indicators. AI recommendation engine generates specific actions per sector: when to irrigate, where pest pressure is building, which areas need fertilizer.

Key Outcomes

• Sector-level health monitoring replaces daily manual field walks
• Critical alerts surface irrigation failures before crop damage occurs
• AI recommendations reduce agronomist consultation costs for routine decisions
• Drone fleet management consolidates mission tracking and battery status
• Farm overview gives investors a single-number summary of operations

The Problem

Licensed land surveyors in Uganda juggle desktop GIS, standalone GPS receivers, manual field notebooks, and separate CAD apps to complete a single job. Data transfer is error-prone. Reports are compiled manually after the fact.

What We Built

A unified land surveying platform connecting field data capture, project management, geospatial computation, and multi-format export into one system — accessible on desktop and in the field.

How We Solved It

Project-centric model with point database, layer set, and measurement records per survey. Map canvas supports Point, Line, Polygon, Measure, Move, Rotate, and Export operations. Layer Manager organizes Boundaries, Survey Points, Contours, Roads, and Rivers independently. GNSS/GPS integration displays RTK Fixed accuracy to ±0.02 m in real time. Export engine produces Survey Reports in PDF, CAD in DWG, GIS in SHP, and Coordinate Sheets in CSV from the same dataset.

Key Outcomes

• Single system replaces GPS receiver + GIS software + CAD tool workflow
• RTK-grade accuracy display gives surveyors field confidence without secondary verification
• Multi-format export delivers all deliverables from one source dataset
• Project management layer keeps office and field teams on the same records
• Offline-first architecture keeps field teams operational without connectivity

The Problem

Ugandans consume news from dozens of fragmented online sources — local outlets, government portals, NGO bulletins, and social media. There is no single destination that aggregates credible Ugandan news, filters misinformation, surfaces relevant job opportunities, and presents everything in a format accessible to users whose first language is Luganda or another local language. Existing news apps are built for global audiences and do not understand the local context, geography, or information needs of a Ugandan reader.

What We Built

A Uganda-focused news and job posting platform that aggregates content from verified local sources, uses AI to personalize the feed per user, flags misinformation before it surfaces, and delivers audio news in Luganda for users who prefer to listen rather than read.

How We Solved It

Content is sourced via RSS feeds from established Ugandan news outlets and enriched with climate and disaster context from the DisasterMap API — so a story about floods in Bududa automatically surfaces with up-to-date field data behind it. A DeepSeek verification layer reviews incoming articles before they are served. Personalization is powered by Neo4j AuraDB, building a graph of user interests and reading history. The job board targets Uganda and East Africa specifically. Luganda text-to-speech is delivered via ElevenLabs.

Key Outcomes

• Single destination for credible Ugandan news replaces fragmented multi-source browsing
• AI verification layer reduces misinformation exposure before content reaches users
• Personalized feed via Neo4j graph matching surfaces relevant stories without manual curation
• Luganda TTS makes quality news accessible to Luganda-first users
• Integrated job board connects readers to East Africa-specific employment opportunities
• DisasterMap enrichment adds real-time climate context to breaking news stories

The Problem

Rural health clinics in Uganda serve large populations with limited doctors. Patient records are kept in paper files that are easily lost or damaged. Patients travel long distances for follow-up visits that could be handled remotely. Specialist consultations require referrals to distant hospitals with no digital handoff — the referred patient arrives with a handwritten letter or nothing at all.

What We Built

A telemedicine and electronic health records platform that gives rural clinics digital patient records, secure provider-to-provider referrals, and virtual consultation capabilities — designed for low-bandwidth environments.

How We Solved It

Patient records are stored digitally with offline-first sync so clinics without reliable internet can continue operating. The referral system creates structured digital handoffs between facilities with attached patient history. Virtual consultations use an adaptive video/audio system that adjusts quality based on available bandwidth. SMS appointment reminders reduce no-show rates. Lab results are uploaded and linked to the patient record immediately.

Key Outcomes

• Digital patient records eliminate lost paper files and transcription errors
• Structured digital referrals replace handwritten referral notes
• Virtual consultations reduce patient travel costs for follow-up visits
• Offline-first design keeps rural clinics operational without reliable internet
• SMS reminders reduce missed appointments in communities with phone access