Suppliers
What to ask an AI supplier before you buy
Do not accept a promise that a product stops prompt injection. Ask how the supplier reduces it, what the tool can reach and how you will see what it did.
Last reviewed against NCSC, Department for Science, Innovation and Technology and OWASP GenAI Security Project. Sources
Questions
Six to put in writing
Reach
What data can the tool see and what actions can it take? The NCSC says to know what access agentic tools have [3].
Privileges
How are privileges reduced when the tool processes content from outside your organisation [1]?
Controls
Does it separate untrusted content, validate output and require approval for high-risk actions [9]?
Supply chain
How does the supplier secure its own supply chain? The code of practice makes this its seventh principle [6].
Accountability
Who is accountable when something fails? The NCSC asks who is accountable for AI security and what the worst-case failure would look like [2].
Responsibility
Not only the user's
The NCSC's view is that the job of using AI safely should not land on individual users of AI products [2]. Reflect that in contracts and in what you expect from vendors, so the supplier carries its share.
Evidence
What a good answer looks like
A good answer is specific. It names the controls, shows where each one sits in the product and says what is left over. Be wary of replies that rely on a single filter or on assurances with no description of the design [1].
Keep the answers
Record each answer with a date. If the product or its permissions change, ask again [3].
Keep reading
Read next
What it is
What AI security means: securing AI systems and using AI safely in defence, the main weaknesses, and the UK guidance that applies to your organisation.
Read the guide
OWASP list
The OWASP Top 10 for LLM applications 2025: all ten risks listed, which matter most for UK organisations, and how to use the list in a review.
Read the guide
Questions
Sourced answers on AI security: what it is, whether prompt injection can be fixed, shadow AI, the UK code of practice and what to do first.
Read the guideHow CyPro can help
Need a hand planning for this?
Understanding the change is one job. Working out what it means for your systems, your suppliers and your roadmap is several more. CyPro's consultants can take any of them on.
Fractional CISO for the board
A senior security lead who advises the board on what to fix first, what to accept and what to tell customers, and owns the decision record.
More on cypro.co.uk
Hardening and segmentation
Our consultants put the controls in place that limit the damage when something is missed: firewall rules, network isolation and tighter admin access.
More on cypro.co.uk
Penetration testing
A focused test of your systems and the controls around them, so you know they hold before an attacker finds out for you.
More on cypro.co.uk
Vulnerability scanning
Find every internet-facing system you run, including the forgotten ones, and see which are still on a vulnerable build.
See the scanning service
24/7 security monitoring
Our SOC analysts watch your systems day and night for the signs of an attack in progress: unexpected processes, new outbound connections, logins that do not fit.
More on cypro.co.uk
Incident response
If you may already be compromised, our responders contain it, preserve the evidence, find out what was touched and get you running again safely.
More on cypro.co.uk
Plan ahead
Want a plan for your organisation?
AI security starts with knowing which tools are in use, who owns them and what they can reach. CyPro can help set the governance, accountability and controls around that.