Clients started asking me about my use of AI years ago, shortly after ChatGPT emerged and people started experimenting with AI-generated output. It started with “We don’t want AI, ever” and has shifted over time to “We use AI to speed up parts of our work.”
Today, I’m always prepared for the discussion with a potential client. It does no good to hedge or dance around the question. Ultimately, your stance and the client’s stance on AI have to be aligned, or the relationship won’t work.
The question is coming up whether you raise it or not
On August 14, Anthropic said it would start embedding an invisible watermark in the text Claude generates to comply with the EU AI Act (but the output will apply worldwide). A few weeks before that, Substack began giving readers the option to run posts through an AI detector built by Pangram.
AI detectors admittedly have problems. Anthropic admits that the watermark will only indicate that text was processed by Claude, not generated by Claude. Many detectors come up with false positives, which can turn into tense situations.
But the point is that people want to understand if AI was used during the creation process. They may have feelings about AI-generated output that fall on both sides of the debate.
All of these tools are aimed at improving transparency, and that’s what you need to consider with clients.
Decide where your line is
You can’t disclose your stance on AI if you don’t have one. You’ve got to set the boundaries of your AI use in your business, whether it’s “use AI as a supplement” or “don’t use AI at all.” Your use of AI may also vary depending on whether it’s client work or for your internal purposes.
Clients will care about the distinction between fully AI-generated work, AI-assisted work, or prefer that you not use AI at all. AI-assisted means a human is involved at every step. AI-generated means no one is. When a client hears “AI,” they may assume the second one, so the specifics of what you do might matter a lot, depending on the client.
Where you land—your “AI type”—is up to you. Some solopreneurs advertise heavy AI use as a feature because they can offer faster turnaround or higher volume—and that’s a legitimate business model for some clients who want it. But other clients may have an opposite expectation.
Be upfront and put it in writing
At a minimum, you should be prepared to discuss your use of AI during your initial calls with clients. You may even want to (or clients expect you to) include clauses in your agreement about your use or non-use of AI.
For example, your agreement might read: “All work delivered under this agreement is original work created by the Contractor. [Additional statement if you use AI assistance in any preparatory steps].”
If your contract hasn’t been updated in several years, you may want to consult with a lawyer about whether AI clauses would be appropriate.
If a client thinks you used AI
A fellow solopreneur recently had a client request that he make changes to something he’d written. The client wasn’t accusing him of using AI, but the content had some “AI tells” (words and phrasing preferences), and the client wanted to eliminate even the appearance of AI writing.
Even if you don’t use AI in any of your work, you can expect that — at some point — a client will ask. It’s a question on nearly everyone’s mind because there’s no single method of determining AI-generated content at the moment.
It starts with the client’s tone. Is it a question, a conversation, or an accusation? If it’s a question or conversation, you may need to clarify your stance on AI if that didn’t happen upfront.
An accusation is different because there’s either a gap in expectations (if you’re using AI) or a lack of trust.
At that point, you should provide evidence to support your response, such as:
- Version history in the document
- Any research notes you took
- How you prepared the deliverable—your “human oversight”
Because AI detectors aren’t perfect, avoiding a false positive is like a game of whack-a-mole. Chances are, this will come up at some point, and you’ll need to be prepared to respond.
The value of transparency
Some solopreneurs are very open about their use of AI. They promote it right on their LinkedIn profiles or websites. Others wait until the conversation comes up with a potential client, and that’s also fine.
Your approach to AI may also change over time. Mine certainly has, as the mode capabilities have improved, and I expect it to move again. You should commit to telling clients how you use AI, both now and in the future, since transparency serves both you and the client.
