Service — AI and Machine Learning
AI Consulting for Public Sector
We design, build, and validate AI systems for UK central government, devolved administrations, and arm's-length bodies. Our delivery teams combine academic rigour with production-grade engineering — all aligned to GDS Service Standard and available via Crown Commercial Service framework RM6200.
What We Deliver
"Public sector AI must be explainable, auditable, and aligned with GDS standards from day one — not retrofitted for compliance after deployment."
— Dr Stylianos Kampakis, Managing Director, Tesseract Academy
Tesseract Academy provides end-to-end AI consulting from problem scoping through to live deployment. Our engagements typically begin with a structured Discovery phase — mapping data availability, identifying algorithmic risk, and defining success metrics before a single model is trained. This approach reduces rework and aligns technical outputs with policy intent from day one.
Our core capabilities span supervised and unsupervised machine learning, large language model (LLM) integration, retrieval-augmented generation (RAG) pipelines, named entity recognition (NER), ontology generation, and digital twin development. We have delivered production ML systems for land valuation, ontology extension for the National Digital Twin Programme, and cybersecurity simulation research in collaboration with the Alan Turing Institute. Our work informs Office for National Statistics (ONS) data interoperability initiatives and NHS England digital infrastructure programmes.
All AI systems we develop are accompanied by bias auditing, Data Protection Impact Assessments (DPIA), and where required, Algorithmic Impact Assessments (AIA) in line withEU AI Act obligations and UK government guidance on algorithmic transparency. In 2024, our open-source AI governance platform catalogued 48 governance tools covering EU AI Act, NIST AI RMF, and ISO 42001 compliance requirements reviewed against Department for Science Innovation and Technology (DSIT) standards.
We are a Cyber Essentials certified micro-enterprise with public liability insurance of £2 million, employers' liability of £10 million, and professional indemnity of £5 million. Our PPON is PWJP-6874-MXDJ. We operate as an SME under Crown Commercial Service (CCS) RM6200 (Management Consultancy Three) and are eligible for direct award for contracts below £10,000. Our procurement credentials are registered with the Cabinet Office supplier register.
Service Comparison
| Capability | Tesseract Academy | Large Systems Integrator | In-House Team |
|---|---|---|---|
| GDS-aligned delivery | Yes — Discovery to Live | Varies by team | Depends on capability |
| AI governance (EU AI Act) | Integrated — 48 governance tools | Additional cost | Rarely available |
| Research-backed approach | Academic + practical methods | Primarily commercial | Limited research capacity |
| SME pricing flexibility | Yes — micro-enterprise rates | Premium enterprise pricing | Fixed headcount cost |
| CCS framework availability | RM6200 | Multiple frameworks | Not applicable |
| Open-source deliverables | Default where permitted | Rare — IP retention common | Possible |
Case Studies
Welsh Government — Machine Learning for Land Valuation
ML-Assisted Land Valuation for National Tax Policy
Commissioned by Welsh Government (2025-2026) to evaluate five land valuation methodologies for the proposed Land Value Tax. Our team applied machine learning models to Land Registry data across Welsh local authorities, benchmarked against conventional valuation and formula-based approaches. Statistical outputs were validated against international comparators and presented to Welsh Government officials. Findings were subsequently cited in Senedd committee proceedings and published on GOV.WALES. The methodology received commendation from HM Treasury evaluators and is referenced in MHCLG guidance on land data standards.
5
Valuation methodologies evaluated
22
Welsh local authorities covered
Senedd
Cited in committee proceedings
National Digital Twin Programme (NDTP) — Ontology Generation
AI-Powered Ontology Extension for National Infrastructure
In collaboration with NDTP (Department for Business and Trade, 2024-2025), we developed an open-source AI tool that automates ontology generation and extension for the National Digital Twin Programme. The tool combines data profiling, Named Entity Recognition (NER), and large language models to extract and generate ontology entities from multiple data formats. The tool is publicly available on GitHub and supports interoperability across UK infrastructure data systems. The project aligns with the UK AI Safety Institute's principles for responsible AI development and has been noted by Skills England in discussions on AI tooling for the public sector workforce. NESTA and Innovate UK have highlighted open-source ontology tooling as critical to the UK's AI infrastructure strategy. The National Audit Office has identified interoperability tooling of this kind as a key value-for-money lever in digital transformation programmes.
3
AI techniques combined (NER, LLM, data profiling)
Open
Source — published on GitHub
NDTP
Dept for Business and Trade collaboration
How to Commission This Service
AI consulting from Tesseract Academy can be commissioned through the following routes:
- 1
CCS RM6200 — Management Consultancy Three
The primary route for AI strategy, model development, and technical assurance. Suitable for engagements from Discovery through to Live phase. Further competition required for contracts above £10,000.
- 2
Direct Award (below threshold)
Contracts below £10,000 may be awarded directly without a further competition, subject to framework rules. Suitable for rapid feasibility studies and proof-of-concept engagements.
- 3
Find a Tender Service (FaTS) — above threshold
For contracts above OJEU replacement thresholds (approximately £213,000 for central government services), full advertised competition via FaTS applies under the Procurement Act 2023.
To begin a conversation, contact us at fabio@thetesseractacademy.com with a brief description of your requirement, data environment, and indicative timeline.
