The full answer
As an AI-first company and an OpenAI Select Partner, we build with models from OpenAI, Anthropic's Claude, Google's Gemini and open models such as Llama. For client products, we use enterprise or API terms where providers do not use your data to train their models, and choose regional hosting where data residency matters.
Data minimisation applies to AI too. We send models only the data a task needs, mask or remove personal identifiers where possible, and in retrieval systems ensure users can only retrieve documents they are allowed to see.
AI features introduce new risks, such as prompt injection, where text in a document or message tries to override instructions. We design guardrails, validate model outputs before they trigger actions, and keep a human in the loop for decisions with real consequences.
Where an open model running in your own cloud is a better fit for sensitive data, we can deploy and fine-tune it there. Our AI engineering team will recommend the right balance of capability, cost and control during discovery.