AI-DPI Sandbox
DSN's AI-DPI Sandbox, Product Design and Technical Support project – a managed cloud environment and residency that helped Nigerian startups build AI on Digital Public Infrastructure.
The AI-DPI Sandbox – Product Design and Technical Support project is conducted by CcHub, DSN and DIAL, funded by Gates Foundation to accelerate responsible, inclusive, and scalable adoption of Artificial Intelligence (AI) and Nigeria's Digital Public Infrastructure (DPI) by early-stage startups and innovators. It is the direct antecedent to NAISH: it proved the model of pairing shared, managed infrastructure with hands-on technical support so that innovators can build, test, and deploy AI without carrying the full integration burden themselves.
The problem it addressed
While AI and DPI platforms are increasingly available, startups face high integration costs, fragmented vendor ecosystems, regulatory complexity, and limited technical capacity to adopt these systems effectively. The sandbox was designed as a market-enablement and capacity-building intervention, not a standalone product – an intermediary layer between startups and multiple vendors (AI model providers, telecommunications platforms, DPI service providers) that reduced time-to-market, lowered cost barriers, and de-risked early experimentation.
What the sandbox provided
A fully managed, secure, API-based sandbox environment unifying:
- AI services – managed LLM and GPU access to accelerate experimentation.
- Nigerian DPI platforms – a mini "Nigerian DPI stack" exposing DPI (e.g., NIN, BVN, Liveness), open-source tools (e.g. DHIS2, PostgreSQL), APIs (e.g. Maps), and apps (e.g. QR code, notifications).
- Network services – SMS, USSD, IVR, and voice channels.
Focus domains: Health, Agriculture, Education, and Financial Inclusion.
Embedded technical support
Beyond infrastructure, innovators received hands-on support:
- Best-practice solution guides – a curated archive of relevant AI + DPI use cases from around the world to accelerate development.
- Solution testing – focus groups of testers for iterative refinement before full-scale implementation.
- Open-source integration support – tools such as DHIS2, Python, and Apache.
- Development platform & training GPUs – cloud infrastructure / GPUs (Azure, AWS, GCP) for solution development and model training.
- Expert access & residency – a 2-week intensive residency plus ongoing expert sessions on product design, AI, DPI, and integrating AI into DPI.
Residency programme
A structured 12-week residency took startups from idea to market-ready:
| Phase | Weeks | Focus |
|---|---|---|
| 1 – Orientation & Prototyping | 1–4 | Problem framing, data analysis, use case, chosen stack, MVP roadmap |
| 2 – Iteration & MVP | 5–8 | Feature prioritisation, sprint planning, TDD, accessibility, compliance, usability testing |
| 3 – Testing, Deployment & Growth | 9–12+ | Live user testing, load / security testing, soft launch, investor readiness |
Participating startups
Nine startups reached pilot-ready solutions:
| Startup | Solution |
|---|---|
| MyItura | MediLoan |
| Clafiya | Clafiya |
| UHC Tech | EasyCover |
| Fertitude | Fertitude |
| Flolog Pharma | MomCare |
| Alajo | Alajo App |
| XChangeBox | Kidashi |
| Evet Africa | Agrocist |
| Eight Medical | Maternal Lite |
Results & lessons
The project demonstrated that a managed sandbox model coupled with sustained technical support materially improves startup readiness, reduces integration risk, and accelerates time-to-market – validating DPI as a practical enabler of inclusive digital services once access barriers are removed.
The headline lesson: the readiness and proximity of technical support is a catalyst for adoption. Positioning support as an embedded, ongoing function rather than ad-hoc troubleshooting reduced friction at the critical points in each team's development cycle.
Relevance to NAISH
The sandbox is the conceptual and operational predecessor of NAISH. It proved demand for shared infrastructure + embedded support; NAISH supplies the missing piece – a sovereign, high-end GPU node – so that the next generation of this model runs on national compute rather than rented cloud. This documentation site itself is modelled on the DSN AI-DPI Sandbox knowledge base.
Project artifacts
Code base
Archive Overview
Gates-funded and related development work relevant to NAISH – the prior deployments, platforms, and use cases the Hub builds on.
GovLLMiner / AIKano
DSN's Government LLMiner – an offline-first, governance-tuned LLM system (piloted as AIKano for Kano State) that turns unstructured government data into evidence for decision-making.