Frequently asked questions.

Senior people. Our leadership team includes former board directors and CTOs with more than 25 years of experience across technology, telecommunications, education, healthcare, construction and other sectors, supported by PhD-qualified AI and machine learning researchers. You will meet the people doing the work before the engagement starts.

Mostly mid-market and enterprise organisations, plus well-funded scale-ups. What matters most is a clear problem and a committed sponsor.

Discovery engagements can usually begin within two to three weeks of agreement.

We are remote-first but regularly work on-site for workshops, discovery and launches.

Small, senior teams. You will meet the people doing the work before the engagement starts.

Yes. Our AI Compliance, Testing & Auditing service tests AI you have built or bought for accuracy, fairness, robustness, privacy and security, assesses it against the regulations and standards that apply to you, and delivers a board-ready audit report with a practical remediation plan.

Yes. Our AI Safeguards, SOPs & Guardrails service gives you a clear AI policy, role-based standard operating procedures, data-handling guardrails, an approved-tool register and staff training, and keeps them current as AI and regulation evolve.

In your environment wherever possible. When we need to process data ourselves, it stays in encrypted, access-controlled environments in your chosen region.

Never without your explicit agreement. We use enterprise model agreements and self-hosted options that do not retain or train on your data.

Yes, as standard, before any data is shared.

We complete security questionnaires routinely and can share our policies and controls on request.

Fixed-price for well-defined phases, time-and-materials for exploratory work, or a monthly retainer for ongoing partnership. See our Pricing page for details.

You own the code, models and deliverables we create for you.

Our smallest engagement is a one-week Evaluation. If you want to see something working, a two-week Proof of Concept delivers a demo of one use case.

Our PhD-qualified AI and machine learning researchers run continuous R&D: evaluating new models, techniques, risks and regulatory developments as they emerge. Those findings flow directly into our client recommendations, safeguards and delivery practices.

Yes. We design, configure and tune guardrail platforms including Google Cloud Model Armor, Amazon Bedrock Guardrails, Azure AI Content Safety and open-source options such as NVIDIA NeMo Guardrails, so every prompt and response in your AI applications is screened for prompt injection, jailbreaks, sensitive data and harmful content.

Yes. Our PHP Security & Vulnerability Remediation service helps organisations, including banks and other regulated institutions, find and fix exposed PHP vulnerabilities, upgrade unsupported PHP versions and harden their applications, using AI-assisted code analysis to move faster.

We are model-agnostic and choose based on evaluation results, cost, latency and data requirements.

Yes. Integration with existing data platforms, CRMs, ERPs and identity providers is a core part of what we do.

Often. We use mature open-source components where they reduce cost and lock-in, and commercial services where they add clear value.