AI tools at work, with guardrails.

Your team already uses AI tools, and some of them are connected to your code, documents and customer data. We help you decide which tools are approved, what they can reach and what staff can share with them.

The problem

  • AI tools in use with no approved list
  • Coding assistants and agents connected to code or SaaS apps with broad access
  • Nothing written down about what can be pasted into an AI tool
  • Former staff whose AI accounts stay open

Customers and investors have started to ask about this in security questionnaires. A short, honest answer is easier to give when the setup is in place.

What we set up

An AI use review

We find which AI tools your team uses, what they can reach and where the risk sits. You get one page: tools found, risks ranked, fixes in order. It is part of the Scale Readiness Assessment.

Guardrails pack

A policy template, an approved-tool list and a short session for the team on safe prompting: what to share, what not to, and why.

Access setup

Single sign-on and MFA on AI tools, SaaS connections placed behind a gateway, and development sandboxes for engineering teams.

Leavers and reviews

Offboarding that also removes AI access, and regular access reviews that include AI tools. Included in Review-Ready.

Done before

We have rolled out Claude at a fast-growing deep-tech company, with written policies, safe-prompting training, SaaS connections behind a gateway, a compliance dashboard, development sandboxes and access models. The same approach, sized for a team of 10 to 60 people, is what we offer.

Where it stops

We work at the workplace layer. Production cloud and code review stay with your engineers. We describe what we review and set up, and we don't claim a company is "AI-safe".

Book a free 30-minute call

Thirty minutes, free, with a written summary afterwards.