navigating a grey area and curious how others have handled it
we use AI across most of our content work now. the output is reviewed, edited, and approved by humans before it goes anywhere. but our contracts with most clients were written before AI was relevant and don’t mention it either way.
clients haven’t asked. the work quality is high. but i’m increasingly uncomfortable with the silence because i think if they knew, some of them would care. not necessarily object, but care.
has anyone proactively disclosed and how did it land? and has anyone had a client discover it without prior disclosure and how did that go?
i’ve had both happen. proactive disclosure: about half of clients respond with mild interest or mild relief, one responded by adding AI requirements to the next contract renewal. nobody ended the relationship.
undisclosed discovery: one client found out through an internal conversation. they weren’t angry about the AI use, they were bothered about not knowing. that distinction mattered. the trust issue wasn’t the tool, it was the omission
from a brand risk standpoint the silence creates more exposure than disclosure does. if a client discovers it themselves, it becomes a news story internally regardless of the quality of the work.
i added a standard AI use paragraph to all new contracts two years ago. it clarifies we use AI as a drafting tool, all content is human-reviewed and approved, and quality standards are unchanged. nobody has objected. a few clients have asked follow-up questions, all manageable
contracts being silent on AI is the norm right now. most clients haven’t thought about it either.
my approach is to disclose in a workflow document rather than the contract. frame it as “here’s how we produce your content” rather than “here’s a legal disclosure.” that framing keeps it professional rather than defensive and it’s easier to update as your process changes
had a client discover it. they’d run a piece through aidetector.ac out of curiosity and it came back flagged. they weren’t angry but they did ask direct questions.
what made it fine: we had a good relationship, the work had performed well, and i could explain the process clearly. what would have made it worse: if the work had also been mediocre, or if i got defensive. the AI wasn’t the issue. the lack of expectation-setting was
the B2B context matters here. enterprise clients have procurement and legal involved. if AI use is discovered during a vendor review or RFP process for a contract renewal and you haven’t disclosed, that’s a different problem than a client casually noticing.
worth getting ahead of it proactively for any client where the relationship matters long-term. a brief addendum to the scope of work is lower stakes than a contract renegotiation