How to explain your automation
to clients.

The distinction that matters.

There is a difference between AI that makes a decision about a client and AI that does the administrative work around a decision you made.

If your automation drafts a follow-up email based on notes you reviewed and approved, the professional judgement is yours. The typing was automated. That is not ethically different from a lawyer using a template, a consultant using a spreadsheet model, or a therapist using a standardised intake form. Tools are not disclosable by default. What you decide and take responsibility for is what matters.

If your automation is involved in decisions that affect the client — prioritising their case, making recommendations, scoring their application — that is different. That is where disclosure becomes a more serious question, and in some regulated professions, a legal one.

Most business automation builds for small business owners fall in the first category. Be honest with yourself about which side yours sits on.

Three client situations and what to say in each.

• A client compliments the quality of something your automation produced. Accept the compliment. You reviewed it. You sent it. It is your work. If they press — "how do you have time for this?" — something like "I've set up a system that handles the first draft" is accurate and invites no further anxiety.
• A client asks if you use AI. Answer directly. "Yes, for the administrative parts — drafting, notes, scheduling. The advice and judgement are mine." Most clients asking this question are not looking for a reason to leave. They are trying to understand what they are getting. A direct answer respects that.
• A client objects to the use of AI. This is rare, but it happens. The right response is to find out what specifically concerns them. Usually it is one of three things: their data being used to train AI models (ask your builder whether this is the case and be ready with a factual answer), a feeling that they are getting less of your attention (address this directly — explain what you review and what you decide), or a general unease they cannot quite name (that one is harder and worth a proper conversation).

The proactive disclosure question.

Some owners decide to mention their automation before any client asks. A line in the engagement letter, a note in the welcome email. There are good reasons to do this and good reasons not to.

The case for proactive disclosure: it removes the awkwardness of being asked and having to explain retroactively. It signals transparency. And for clients in certain industries — legal, medical, financial — it may already be expected or required.

The case against: most clients do not think about what software you use any more than they think about which brand of phone you make calls on. Proactive disclosure of every tool in your business is not a standard expectation. Disclosing your AI automation specifically, while not disclosing your calendar software or your accounting app, implicitly flags it as something unusual or concerning when for you it is not.

The middle path: include a brief, plain-language note in your standard engagement or onboarding materials — not as a confession but as information. "I use software to handle the administrative parts of my practice, including drafting and note-taking. All advice and decisions are my own." That sentence does the job without drawing disproportionate attention.

What to do if your profession has specific rules.

Some professions — lawyers, doctors, accountants, immigration advisers — have regulatory guidance on AI. Most of that guidance is still evolving. The absence of a specific prohibition is not permission, and the presence of guidance does not always mean enforcement.

The safest position is to check with your professional body before your automation is running, not after. Ask two questions: does using AI for administrative tasks require disclosure to clients, and does it require any additional documentation in client files? The answers are usually short and practical.

Nina · Independent immigration adviser — she emailed her professional body with a two-paragraph description of what her automation does. The response came back in four days: document review and note-drafting were fine, any AI involvement in assessments or recommendations needed a client disclosure. She knew exactly where the line was before she turned the automation on.

What not to do.

• Deny it when asked directly.

  If a client asks "did AI write this?" and the answer is "a draft, yes, and I reviewed and edited it" — say that. Saying no is a lie with a short shelf life and a long tail of problems if it comes out later.

• Over-explain unprompted.

  A four-paragraph explanation of your AI infrastructure in a client email, when nobody asked, creates anxiety where none existed. Answer the question in front of you, not the one you are afraid of.

• Use your automation on a client who explicitly said they don't want AI involved.

  If a client makes this a condition of working with you, honour it or decline the work. Using it anyway and not mentioning it is the version of this that ends badly.

• Confuse "AI drafted this" with "AI decided this."

  The first is a tool. The second is a delegation of your professional responsibility. Keep the distinction clear in your own mind before you explain it to anyone else.