WorkshopScoped individually
Agent Architecture Workshop
A technical session in which we design the agents, their tools, the limits of their access and the points where a person approves.
Domain H · AI systems
AI Governance, Security & Assurance
We put safeguards around AI systems that a business, its customers and its auditors can check: who may do what, which data reaches which model, how quality is measured and what happens when something goes wrong. This is technical governance, not legal advice.
Status:Custom engagementTaken on as a scoped project after an assessment; every engagement is quoted individually.
Who it is for
- Companies putting AI in front of customers or into decisions.
- Teams that must answer data-protection or customer security questionnaires.
- Organisations preparing for the obligations of the EU AI Act, together with their own legal counsel.
AI Governance, Security & Assurance
Problems it addresses
AI deployed without clear accountability.
Customer or auditor questions nobody can answer.
Personal data sent to models without terms.
Agents with broad permissions.
No evidence of quality.
AI Governance, Security & Assurance
Service families and capabilities
The status shows what we offer today. Pricing is set once the scope is clear, in a written quote.
AI risk and use-case review
Custom engagementClassification of use cases by impact, approval requirements, documentation duties — with your legal counsel.
Approval and permission design
Custom engagementDeterministic authorisation around model output, human approval matrices, audit trails. Part of delivery in AI agent projects.
Data protection engineering
Custom engagementData minimisation, redaction, retention, residency, DPA requirements for providers.
Security engineering for AI
Custom engagementPrompt-injection defences, tool permission boundaries, secret handling, threat models.
Evaluation and assurance
Custom engagementTest sets, red-team style probing within agreed scope, evaluation reports, regression gates.
Policy and documentation
Custom engagementAI usage policy, model and data inventory, incident procedure.
AI Governance, Security & Assurance
Workshops and assessments
Consultant-led formats with a clear scope that we usually start with.
ReviewScoped individually
Responsible AI Governance Review
A review of the rules, roles and controls for your AI systems: risk, human approval, data, evaluation, documentation and monitoring.
What you receive and keep
- A risk register per AI system.
- An approval matrix.
- A data-flow and processing inventory.
- A threat model.
- An evaluation report.
- Policy drafts for legal review.
- Implemented controls: redaction, permission checks, audit log.
How we engage
Assessment
A fixed-scope review with written findings, risks and a prioritised plan.
Workshop
Half a day to two days with your team; decisions and a written summary.
Architecture
Target design, decision records, interfaces, security model and an estimate of the work.
Implementation
Built and tested increments with acceptance criteria and handover documentation.
Managed service
Through Managed AI & Platform Operations
Pricing is set once the scope is clear and is provided in a written quote.
What we need from you
- An inventory of AI systems and data flows, or access to build one.
- Your legal or data-protection contact.
- Agreement on the scope of any adversarial testing, in writing.
What we do not do or promise
- No legal advice, no certification and no compliance guarantee. Regulatory interpretation stays with your lawyers.
- Penetration testing beyond agreed AI-specific probing is referred to specialised providers.
Related domains
AI Agents & Business Automation
Agents and automated workflows inside your systems — with permission checks and human approval.
AI Applications & Knowledge Systems
AI assistants and features on your own documents and data — with sources, permissions and measured quality.
Managed AI & Platform Operations
Ongoing operation of what has been built: monitoring, maintenance, evaluation, cost and incidents.
AI Strategy & Transformation
Where AI is worth building, in what order and with which safeguards, before you invest in a project.
Related tools
Start with a quick self-assessment before we talk.
- Available now
Architecture
AI solution builder
Compose a solution from agents, integrations and approval points and get an architecture diagram to discuss.
- For:
- Technical leads and product teams
- You get:
- An architecture diagram of the solution with agents, integrations and approval points.
- Time:
- about 5 min
Request an AI system risk review
Describe your task. We will clarify the scope and the engagement model and send a written quote.
- Written quote with a clear scope
- Transparent milestones and timelines
- Your data is used only to reply to you