Skip to content
All posts
Tools11 June 2026 · 5 min read

Free vs private AI: the plan you choose decides who sees your data

Not all AI tools treat your information the same way. The difference between a free chatbot and a business plan is the difference between exposed and protected.

By Fez Yousuf
Two vaults on a dark surface, one open and spilling its contents, one sealed inside a ring of warm gold light

People often ask me whether AI is safe to use in their business. It is the wrong question, and a completely understandable one. AI is not one thing. The same brand can be wildly different depending on the plan you are on, and the plan is what decides who can see your data.

So the better question is: which version of this tool am I using, and what does it do with what I type?

The free tier is the product

Free consumer AI tools are not charities. When something is free, your usage is often part of how the product improves. On many consumer tiers, the text you enter can be retained and used to help train future models unless you dig into the settings and turn it off. That is fine for asking how to word an email. It is not fine for a client’s financial statement.

This is the trap small and medium businesses fall into. The free tool works beautifully, so it becomes the daily habit, and client information starts flowing through a service that was never designed to keep it confidential.

The business tier is a different deal entirely

Step up to the business and enterprise plans and the relationship flips. Claude for Work, ChatGPT Enterprise, and Microsoft 365 Copilot are sold under commercial terms that, by default, do not use your data to train their models. They add controls a business actually needs: how long data is kept, who can access it, and administration over the whole team rather than each person fending for themselves.

You are no longer the product. You are the customer, with a contract that treats your data as yours.

What to look for before any tool touches client data

You do not need to be technical to vet a tool. Ask four things. Does it train on my data by default, and can I turn that off in writing, not just a toggle. Can I control how long my data is kept. Do I get admin controls across my team. And where is my data stored and processed. If a tool cannot answer these clearly, it is not ready for your client information.

You probably do not need to self-host

There is a myth that real privacy means running your own AI model on your own server. For the vast majority of small and medium businesses, that is expensive, slow to maintain, and unnecessary, because the business tiers already give you the contractual privacy you need. Self-hosting only earns its keep in rare cases with strict data-residency or regulatory rules, and even then there are Australian-hosted cloud options before you ever buy hardware. Do not let the myth either scare you off AI or push you into something far heavier than you need.

The honest summary: you can get the speed of modern AI and keep client data private at the same time. It mostly comes down to being on the right plan and having it set up properly, which is the quiet, unglamorous work that actually keeps a firm safe.

Want this sorted for your business?

An AI Readiness and Security Assessment is a short, fixed-scope look at where AI fits and where your data is exposed, with a plain plan to fix it.

Book a free discovery call

Keep reading, or get a plan.

Browse more plain-spoken notes, or book a free call to talk about your business.