Max Modlin · Founder · 14 October 2026It can be, if the consultancy can answer five questions clearly: where your data is processed, whether it is used to train anyone’s model, what access the system has, how its actions are logged, and who owns the result. If those answers are vague, the data is not as safe as it should be.
Why this question comes first
AI systems are only useful when they work with real business information: customer orders, supplier invoices, pricing, contracts. That is exactly the information you cannot afford to lose control of. A sensible buyer asks about data before asking about features.
1. Where is the data processed and stored?
Ask which AI services the system uses, where they run, and where any stored data sits. You are looking for specific answers, not reassurance.
2. Is it used to train anyone else’s model?
Business-grade AI services offer terms under which customer data is not used to train their models. Make sure the consultancy uses those services and will confirm it in writing.
3. What access does the system have?
The system should use its own controlled access to your ERP, mailboxes and portals, with only the permissions the job needs. That way it can be audited, limited and switched off cleanly, the same as any member of staff.
4. Is every action traceable?
Every record the system writes should be traceable to the email or document that caused it, and every decision it was unsure about should be visible to a person. If something looks wrong, you should be able to see exactly why it happened.
5. Who owns it afterwards?
The safest arrangement is a system built in your own environment and accounts. The code, the configuration and the data stay with your business, whoever supports it in future.
How we handle it
- We build in your environment and on your accounts, so the system and what it knows stay with you.
- Our systems use the same controlled access your people do, with only the permissions the job needs.
- Every field is validated before it is written, and uncertain cases go to a person.
- Every action is logged and traceable to its source.
Our founder spent a decade building automation inside some of the world’s largest financial institutions, and we hold our systems to that standard. More on how AI connects to your existing systems and how to choose an AI consultancy.
Questions we are asked
Is our data safe with an AI consultancy?
It can be, if the consultancy can show you where your data is processed, who can access it, whether it is used to train anyone else’s model, and how every action the system takes is logged. Ask for those answers in writing before any build starts.
Will our data be used to train AI models?
It should not be. Business-grade AI services offer terms under which your data is not used for training. Ask the consultancy which services they use and to confirm this in writing.
What access does an AI system need to our systems?
Only what the job requires. A well-built system uses its own controlled access with the minimum permissions, so it can be audited and switched off without affecting anyone else.
Who owns the system and the data afterwards?
You should. Ask for the system to be built in your own environment and accounts, so the code, the configuration and the data stay with your business if the relationship ends.
Related guides
How to choose an AI consultancy in the UK
A practical checklist: production proof, who does the work, data and ownership, a small first project, and what happens after go-live.
Read → 15 August 2026 · IntegrationWill AI work with our existing ERP and business systems?
Nothing gets ripped out. How AI sits alongside Microsoft Dynamics 365 (Business Central) and the platforms you already run.
Read →