Software and AI for public organisations
We build AI in the public sector for organisations that serve residents, students and communities with limited people and public money. Our tools prepare files, answer routine questions and connect systems on open standards, and officials, teachers and staff make the decisions.
Industries in public sector
The work, and where software fits.
Municipalities, universities and charities share a pattern: rising demand, tight budgets and a duty to explain every decision. Staff spend much of their time finding documents, answering the same questions and moving data between systems bought from different suppliers.
Software fits where it takes over that preparation work and leaves judgement with people. We build document processing, cited search, service agents and integrations on standards such as the ZGW APIs and LTI, with accessibility, EU hosting and code ownership as starting points.
How we improve public sector.
File preparation for decisions
Case files, Woo requests, exam board requests and grant applications gathered, checked and summarised for the person who decides.
Answers for residents and students
Service agents that answer routine questions from checked sources and hand personal cases to staff with a summary.
Search across archives and regulations
Council decisions, examination regulations and programme reports searchable by question, with every answer linked to its source.
Integrations on open standards
Common Ground, ZGW APIs, LTI and SURFconext used to connect systems and reduce dependence on single suppliers.
Accountable and accessible by design
Decision logs, algorithm register details and WCAG 2.1 AA interfaces delivered as part of each build.
Further reading.
Human-in-the-loop AI agents: design patterns for production
The patterns we use to run AI agents safely in production: approval checkpoints, confidence thresholds, tool permissions, audit trails and fallbacks.
4 min read
How to build a production RAG chatbot: a practical guide
What it takes to turn a promising RAG experiment into a chatbot people trust: ingestion, hybrid search, citations, permissions and evals.
5 min read
How to evaluate RAG: retrieval metrics, faithfulness and golden sets
How to measure a RAG system properly: separate retrieval from answers, check faithfulness claim by claim and build a golden set you can trust.
4 min read
Common questions.
Some public uses are high-risk, such as assessing eligibility for benefits or evaluating students. We identify these early, design for human oversight and documentation, and supply what you need for your algorithm register and internal review.
Yes. We deploy in your own cloud tenant or with a hosting party you choose, and use EU-hosted models. Personal data can be masked before any text reaches a model.
You do. Code, configuration and documentation are delivered to your repositories, so your organisation can maintain it, reuse it or hand it to another supplier.
With one bottleneck that is well understood and has a clear owner, such as Woo requests, student questions or grant intake. Our AI readiness checklist helps you choose and prepare the data.