Security and governance

Your data stays yours. A person stays in charge.

The first question every owner asks is "is this safe?" Here is how we answer it, in writing.

Business accounts only
We set up business versions of AI products, whose terms say customer business data is not used to train their models by default. We read the current terms for each product we recommend, because makers change them, and show you where it is written.
In your name
Every account is yours. We use the access we need for the job and nothing more.
Access removed
When a job ends, our access is removed unless a support agreement keeps it. That protects you, and us.
A person approves
AI drafts; your people check and send. Anything that reaches a customer, moves money or changes records waits for a person.
A record of what AI did
Where it matters, an activity log shows what each AI tool did and who approved it.
Staff rules
A short AI policy that says what staff may and may not paste into AI, explained in plain words in training.
Patient data
Clinics and dental practices: patient data is used only with products that meet patient privacy rules (HIPAA), with the required agreements signed first.
Defense suppliers
We check export-control rules (ITAR) before any technical data moves. Where data must stay in-house, AI can run on your own servers.
Your ideas stay yours
What we learn about your business stays confidential. We reuse our general know-how, never your information, and name you as a client only with your written permission.
AI governance

Rules for AI you can show anyone who asks.

Governance is your company's rules for using AI: what it may be used for, who approves a new tool, what data may go in, how its work is checked and how often the rules are reviewed. Customers, insurers and regulators are starting to ask for proof.

In every project

The basics

  • An AI policy staff can read in five minutes
  • Company AI accounts, and a list of approved tools
  • One named person who approves new tools
  • A person checks anything that reaches a customer, moves money or changes records
  • The rules read again every six months
For regulated and larger companies

The full version

  • An approval step for every new use of AI, with a small review group
  • Accuracy tests that measure how often each AI tool is right
  • An activity log of what each tool did and who approved it
  • Regular checks that the rules are being followed
  • Reports for regulators drafted from the same records

Have a specific concern?

Ask it on the first call. We'll answer it in writing in your Snapshot.

Book a call