INREM Foundation

India

INREM trains community-level Water Quality Champions to test, report, and act on water contamination across 8 Indian states. Since 1994: 13 million people reached, 60,000+ champions trained, and a 22% increase in safe water adoption in key intervention areas, with a target of 1 million champions by 2030. INREM is the Government of India's National Thematic Lead for Water Quality under the Jal Jeevan Mission. JalXChange, their WhatsApp-based RAG advisory bot, is live in Assam with 44,000 students engaged, with IBM as an active tech partner. At Matching Day they want a tech partner to strengthen the RAG model, expand multilingual capability, and scale to 500,000 active users.

Race For The Baltic

Sweden

Race For The Baltic identifies, funds, and scales cost-effective interventions to reduce phosphorus entering the Baltic Sea, running eight programmes across agriculture (GYPREG gypsum treatment), coastal wetlands (BalticReed), dry bulk ports, and municipal stormwater (City Accelerator) in Sweden and Poland since 2013. Methodology validated against HELCOM standards, with a signed Polish Ministry pilot for the Obra canal system. BalticEye AI is a complete interactive prototype ranking 34 agricultural entities by AI-generated risk score on the real Obra canal geography. At Matching Day they want a tech partner to build the production version: ETL pipelines from five Polish government registries, an explainable AI classification model, and multi-tenant architecture for all 16 Polish regional inspectorates.

Socios En Salud Sucursal Peru

Peru

Socios En Salud, Peru's affiliate of Partners In Health, brings modern medical science to low-income communities through two interventions live for 5+ years in Peru: Thinking Healthy (WHO-based perinatal depression approach, PHQ-9 validated) and CASITA (early childhood development, ASQ screening), with findings shared with Peru's Ministry of Health. MiWawa, their digital prototype combining both, has been piloted with around 3,000 individuals in Lima. At Matching Day they want a tech partner to build Phase 1 AI models for dropout prediction and PHQ-9 risk trajectory detection, plus Phase 2 computer vision models analysing caregiver-submitted videos for developmental risk signals, all as decision support for community health workers.

Lebanese Alternative Learning

Lebanon

Lebanese Alternative Learning runs Tabshoura, a free curriculum-aligned digital learning platform live for 5+ years with 68,000+ individuals reached across 230 institutions. Content covers early childhood through middle school in Arabic, English, and French, working online, offline via Tabshoura-in-a-Box, and integrated with UNICEF Madrasati. Tabshoura has become the de facto emergency education platform for 1.2 million+ displaced Lebanese. At Matching Day they want a tech partner to build an AI recommendation engine personalising learning paths, a multilingual chatbot (Arabic, French, English) for 24/7 learner and caregiver support on web and WhatsApp, and rebuild the mobile app as offline-first for low-end devices.

Neurodiversity Foundation

Netherlands

The Neurodiversity Foundation runs a global movement for neuroinclusion through Neurodiversity Pride Day (2.1 million people reached across 58 countries in 2025), policy advocacy (Motion 370 passed in the Dutch parliament), and assistive technology research. They run a global volunteer workforce of 180+ people, 85% of whom are neurodivergent, across 8 distinct roles and 20+ projects. They've built over a dozen custom AI tools in-house, including chatbots, a grant-finder agent, and a Python-based team chart builder. At Matching Day they want a tech partner to build NeuroScale, a unified AI agent for volunteer onboarding combining RAG with LLM orchestration, plus a chapter-in-a-box deployment model for international expansion.

Centrum Edukacji Obywatelskiej (The Center for Citizenship Education)

Poland

The Center for Citizenship Education is Poland's largest educational NGO, running for 31 years. Their two flagship electoral programmes reach 4.6 million citizens per election cycle through Latarnik Wyborczy (Electoral Lighthouse) and 200,000 students per year through Młodzi Głosują, across 10,000 Polish schools. MEWA, their RAG-based electoral chatbot built by Polish tech firm Nomtek on a Python/LlamaIndex/OpenAI stack with RAGAS bias evaluation, cites specific articles of the Electoral Code and has been tested with young voters. At Matching Day they want a tech partner to integrate MEWA into Electoral Lighthouse, expand its knowledge base into political and historical context, and build a parliamentary transcript pipeline.

Kheyti Tech Pvt Ltd

United States

Kheyti deploys affordable protected cultivation infrastructure (the Greenhouse-in-a-Box, 90% cheaper than conventional greenhouses) to smallholder farmers across 8 Indian states. Third-party validated outcomes in year one: 73% net income increase, 58% reduction in crop loss, 70% savings on water and fertiliser. 8,000+ farmers reached and Earthshot Prize 2022 winner. Their WhatsApp advisory chatbot is live in Telugu and Hindi, integrated with Plantix image-based plant health diagnosis and 20,000+ seasons of proprietary farm data. At Matching Day they want a tech partner to push image-based diagnostic accuracy toward 90%+ confidence, expand multilingual flows across crops and geographies, and optimise the human-in-the-loop agronomist review workflow.

Display Europe eG

Austria

Display Europe is a Vienna-based media cooperative distributing independent European journalism across language barriers, with 75+ media outlets connected and 50,000+ individuals reached. Their REPCO vector database, hosted in Austria and GDPR-compliant, ingests journalism from any CMS and translates it into 20+ languages. The Display AI Assistant overlays content on partner publishers' sites while returning traffic to the original publisher rather than retaining it. At Matching Day they want a tech partner to replace commercial LLMs with self-hosted, fine-tuned Small Language Models on sovereign European infrastructure, and build a multilingual semantic search and personalised content recommendation engine.

MedNTech

United Stated

MedNTech, led by an all-women team and a WHO Civil Society Commission member, built CerviScanner, an AI-powered smartphone app supporting frontline health workers during Visual Inspection with Acetic Acid cervical cancer screening, targeting WHO 90-70-90 elimination goals. Piloted at Kicukiro Health Centre, Rwanda in 2025: 260 women screened, 84.5% AI diagnostic accuracy and 92.2% specificity on a 251-woman validated subset. The app runs an EfficientNet-B0 CNN trained on the IARC dataset on standard Android phones with no additional hardware. At Matching Day they want a tech partner to transition CerviScanner to native Android offline-first with Kinyarwanda and Swahili localisation, plus an image preprocessing model and synthetic data generation pipeline.

MedCycle Network

United States

MedCycle Network, a Fast Forward finalist, collects donated surplus medical supplies and equipment from hospitals for free distribution to safety-net clinics serving low-income, uninsured, and Medi-Cal patients. Since July 2024 they've redistributed $1.6M+ in surplus to 25+ clinic partners across California, supported 180,000+ patients, and diverted 25+ tons from landfills, with a verified 3 to 10x multiplier on philanthropic investment. ReadyMedAI's web portal is live on AWS with open-source architecture, with semi-automated matching achieving 90%+ delivery acceptance. At Matching Day they want a tech partner to build the core AI: an agentic matching engine with equity-weighted scoring, predictive demand forecasting, hospital inventory API integrations, and an open-source replication toolkit.

FUNDAMEDIOS

Ecuador

FUNDAMEDIOS defends freedom of expression, press freedom, access to information, and digital rights, and fights disinformation across the Americas, with around 1 million people reached indirectly. Liberties360, their MVP civic intelligence system, is live in Ecuador (press-watch-latam.lovable.app), combining continuous public source monitoring, structured risk classification, human verification, and a public dashboard. Web scrapers for Twitter, YouTube, and media sources are in development. At Matching Day they want a tech partner to automate data collection pipelines, build ML models for risk classification, pattern detection, and alert prioritisation with human verification in the loop, and refactor the backend for multi-country scalability.

CDR

Turkey

The Center for Democracy Research strengthens democracy, civic participation, and accountable governance in emerging democracies through civic technology (Policyful, Referandom, Gov2AI), research, and capacity-building. Their Referandom platform ("Tinder for Parliament") reaches 10 million+ people per month. Policyful, their open-source policy intelligence platform, is in beta in Turkey with web app, cloud infrastructure, data warehouse, APIs, and AI analytics components in place, with a grant secured for expansion into Bulgaria, Poland, and Brazil. At Matching Day they want a tech partner to automate document ingestion and categorisation pipelines, develop RAG-based document analysis, and build multilingual model adaptation for country-specific deployment.

Build Change

United States

Build Change reduces deaths, injuries, and economic losses from housing and school collapses due to earthquakes, extreme weather, and fires. Their BCtap SaaS platform (React Native, Next.js, AWS) has run for 20+ years across 6 countries, making 1.4 million people safer and serving 283,000+ households. Tara, their AI chatbot using RAG on LangChain, is live in the Philippines answering homeowner and MFI loan officer questions in English and Tagalog with source-cited responses. At Matching Day they want a tech partner to evolve Tara into agentic AI connecting to live assessment data, reasoning over each homeowner's construction typology against a Technical Ontology for Informal Buildings to generate personalised step-by-step construction guidance.

Farmers for Forests

India

Farmers For Forests restores biodiversity and strengthens rural livelihoods with agrarian and indigenous communities across India, combining agroforestry implementation with AI-enabled digital monitoring. TreeLens, their drone-based AI MRV system, monitors 4,000 plots across 5,000 acres and 1.7 million trees in 6 Indian states with 14+ partner organisations. The model runs on Meta's Detectron-2, trained on 25,000+ tree crown annotations, achieving 85%+ accuracy on small trees and 92%+ on mature trees, with outputs published on a public dashboard and a free TreeLens web interface. At Matching Day they want a tech partner to automate the full drone processing pipeline (12 to 24 hour turnaround target), build a central GIS data catalog, and add retraining automation for new geographies.

Travel Hands

United Kingdom | Disability access and mobility

Travel Hands pairs visually impaired people with DBS-verified, trained sighted guides for safe outdoor journeys across 22 London boroughs at £3 per journey. 1,500+ journeys completed and 400+ users supported, backed by JP Morgan Chase and the UK Space Agency. The platform runs on a production-grade AWS backend (Elastic Beanstalk, DynamoDB, S3, SNS) with GPS journey tracking, in-app DBS verification, and a real-time customer service dashboard. VIPA, their AI voice chatbot, already lets users book a journey end-to-end by voice. At Matching Day they want a tech partner to fine-tune LLaMA on real journey data, integrate live accessibility sources (Street Manager API, TfL, National Street Gazetteer, OpenStreetMap), and build real-time safety prompts and escalation logic.