AI setup for small teams
A safe way for your team to use AI with customer data
Your staff already paste emails into chatbots. I help you choose the approved tool and account, write the one-page policy, and teach the habits that keep customer data where it belongs.
Prefer to write first? Send a message instead.
Fixed prices, excl. VAT · You own everything · Reply within 24 hours
Why the free chatbot in the next tab is the risk
The wrong account is the usual slip
Regulators report breaches where staff used a free chatbot while the employer paid for a business version with data agreements. Paying for a personal subscription does not turn it into a business product.
Nobody has sorted the data
A published price list and a patient's treatment notes are not the same thing. Without clear zones, people guess, and guessing is how health details, ID scans or NDA material end up in a cloud chatbot.
The policy is missing or unread
Most small teams have no written rule on which tools and accounts are allowed, and your privacy statement may not list the AI providers that receive personal data.
How we get your team onto one safe setup
-
Find out what is really in use
We look at which accounts and tools your team uses today, and at what you already pay for. Microsoft 365 or Google Workspace may already include an assistant covered by your existing data terms.
-
Sort your data into three zones
We split what you handle into public or anonymised, ordinary personal data, and special or secret data, using real examples from your own business.
-
Choose tools and accounts per zone
Each zone gets an approved tool and account type, checked against what the provider states about training, retention and its data processing agreement. For the most sensitive tasks we test a local model on your own documents.
-
Write the one-page policy and teach the habits
You get a short policy people will read, plus a worked example of anonymising a prompt from your own inbox, and a checklist for what to do after a slip.
I read data terms before features, on my own tools
For my own personal assistant I dropped free model endpoints whose terms log prompts, and I made the router stop and ask before any request goes to the cloud. These are my own projects, not client work.
I removed a claim I could not back up
I took an unbacked number off my own AI page and I never call a tool "GDPR compliant". I tell you what a setup reduces, and where the legal decision belongs to a privacy professional.
Where to start
The usual first step is an audit of where your team uses AI today and what each zone of data needs. If you then want a workflow built on top of that, it follows from the same sorting exercise.
What you get
AI Audit & Roadmap · €549
Workflow review ending in a written, prioritised roadmap within 5 working days
- Intake call + workflow review
- Written report in plain language
- Prioritised roadmap with effort estimates
- Delivered within 5 working days
- Credited against any build booked within 60 days
What you get
Custom Solutions · from €2,500
The build step: automations, custom apps and integrations, scoped after the audit
- Workflow automation inside the tools you already use
- Custom apps and internal tools built around how you actually work
- Agent and API integrations between the systems you already pay for
- Model selection: which model, running where, and why
- Scoped in writing before any work starts, with defined success criteria
- A one-page data map: what enters, what leaves, where it is processed
- Documentation, a walkthrough call, and ownership of everything
Common questions
Is ChatGPT safe for customer data?
On a personal account with default settings, no. A business account with a signed data processing agreement can be acceptable for ordinary customer data. Health details, ID documents and anything under an NDA should stay out of cloud chatbots.
Can you make us GDPR compliant?
No tool or consultant can hand you compliance, because compliance is something your business does. I can set up the tools and accounts so that doing it is easier. Reporting decisions are legal ones, and I am not a lawyer.
We already pay for Microsoft 365 or Google Workspace. Is that enough?
Often it is the safest starting point, because the assistant may be covered by the data terms you already have. We check this first, then make sure staff use that account rather than a personal one.
What if someone has already pasted customer data into a chatbot?
Act the same day: find out what went in, delete the conversation, assess the risk to the people involved and record it. Whether a report to the regulator is required is a question for a privacy professional.
Get one approved tool and one page of rules
Tell me which tools your team uses now and what data they handle. I will say what is worth fixing first.
Prefer to write first? Send a message instead.
Fixed prices, excl. VAT · You own everything · Reply within 24 hours