AI Transparency

1. Our AI Principles

• Human oversight remains accountable for every decision that affects a client outcome.
• AI is used to augment senior practitioners, not to replace expert judgement.
• We prioritise accuracy, explainability and reproducibility of AI-assisted outputs.
• Client confidentiality and privacy take precedence over automation convenience.

2. How We Use AI

AI is used as a supporting capability within our software delivery process. Depending on the scope of an engagement, carefully selected tools may assist with requirements analysis, documentation, code and test review, planning, reporting, risk identification and delivery visibility.

These tools do not make final decisions or replace accountable professionals. Their outputs are reviewed in context, and responsibility for quality, delivery and outcomes remains with experienced people.

3. Human-in-the-Loop

AI-generated suggestions are reviewed by qualified engineers before they influence client deliverables. Critical decisions — architecture, security posture, release approvals — remain the responsibility of named humans.

4. Data Handling

We do not use client data to train foundation models. Where third-party AI providers are used, they are selected for their data-handling commitments and configured to disable training on submitted content.

5. Known Limitations

Generative AI systems can produce plausible but incorrect outputs. Our workflows are designed to detect and correct such issues through validation, testing and expert review before delivery.

6. Governance

Our internal AI governance aligns with widely recognised principles for responsible AI use and evolves with regulatory guidance such as the EU AI Act. Clients can request additional information about our controls under NDA.