Templates
Reference solution-architecture templates showing how the NAISH Hub powers AI use cases across public service delivery, government, health, agriculture, and education.
Templates are reference solution architectures. Each shows how a real class of solution is put together, and – most importantly – where the NAISH Hub fits in and what AI it powers. They are starting points a team can adapt: keep the shape, swap in your own models, data, channels, and platforms.
The examples below are illustrative, not prescriptive. A template names a kind of component – a records system, a messaging channel, an identity check – not a specific product. Use whichever tools and platforms suit your project; the Hub is the constant.
What every template shows
- The channels users arrive through (chat, voice, apps).
- The AI layer the Hub powers – models served on Hub GPUs, an agent that reasons over a goal and calls tools (its skills), and retrieval that keeps answers grounded.
- The supporting services and data the solution reads and writes.
- The external dependencies it relies on (identity, payments, messaging, model sources).
The Hub's role is consistent across all of them: it is the sovereign AI core – serving and fine-tuning models on national infrastructure – that these solutions plug into. See GPU / Compute Support.
The templates
Public Service Delivery
A multi-channel citizen-service assistant: an AI agent over records and workflow, with identity and payment dependencies.
Government
A policy and decision-intelligence assistant: document ingestion, grounded retrieval, and dashboards.
Health Coaching
A multi-channel health coach grounded in clinical guidelines, with reminders and a strict safety posture.
Health Records
AI-assisted record digitisation and natural-language lookup over structured records.
Agriculture
A multilingual advisory assistant grounded in agronomy, with weather-aware alerts over IVR / SMS.
Education
A reading and learning tutor: speech scoring plus curriculum-grounded tutoring with progress tracking.
Local Commerce
A selling assistant for local traders: photo-to-listing, price advice, and orders with payments and alerts.
Workloads
How inference, training, and data workloads are allocated across the node under the governed multi-tenant model – and why the allocation pattern matters for capacity.
Public Service Delivery
Reference architecture for a multi-channel citizen-service assistant – an AI agent over records and workflow, with identity and payment dependencies, powered by the NAISH Hub.
