Tier two is everything that does not carry the immediate consequence of a factual error but still carries your professional reputation. This is the content layer: emails, proposals, itinerary introductions, social posts, research summaries, anything where the risk is not that a client is harmed but that your communication does not meet the standard they expect from you.
The editorial review habit is simpler than tier one because it does not require external sources. It requires you to read the output with the same critical eye you would apply to anything leaving your practice, regardless of who or what produced the first draft.
Here is the checklist I use. Five questions, in this order.
Does it sound like me? If it reads as generic AI-generated text, the prompt lacked specificity or the context document needs refinement. Either way, it does not go out until it sounds like my practice, not like a template.
Is it right for this client? Not just correct, but appropriate. The right tone for where they are in the booking process. The right level of detail for what they need at this stage. The right register for the relationship.
Is there anything here I have not confirmed? A factual claim I am taking on trust, a detail I am assuming is current, a reference I have not checked. If there is, it either gets confirmed or removed before the output leaves.
Is it the right length? AI tends toward thoroughness, which in practice often means over-length. If the communication would land better at half the length, edit it to half the length. The client’s attention is a finite resource.
Would I be comfortable if the client knew how this was produced? Not because there is anything wrong with using AI, but because the output should reflect the quality of thought and care that you would want to be associated with regardless of the production method.
Those five questions take less than two minutes to work through. They are the difference between output that leaves your practice and output that represents your practice.
