My team uses different AI tools and nothing matches

I start the morning by opening the latest client brief and immediately see the problem. Three people helped draft it yesterday, and all three used AI. One section is sharp and structured. One reads like a generic blog post. One includes a table with figures I do not recognise. The work is faster than it used to be, but it does not yet feel like one team produced it.

By mid-morning I am comparing the draft against last month's approved example, rewriting headings, and asking people which sources they used. There is a small win: one teammate has clearly found a useful prompt for summarising meeting notes, and that section gives me something to build from. But the rest of the file still needs too much judgement from me before I can send it anywhere.

After lunch, a compliance question lands in my inbox. It is not dramatic, but it is uncomfortable: did anyone paste customer information into an AI tool, and which tool was used? I can ask the team, but I do not have a simple record. Everyone was trying to help. Nobody was trying to be careless. We just never agreed what information was safe, what needed anonymising, or what had to stay out of public tools.

Later, a new hire messages me and asks whether we have an AI template for client work. I nearly send them three old examples and a note saying 'copy this style,' then realise that is exactly how the problem keeps repeating. They need a clear rulebook, not another vague example to interpret.

By the end of the day, the brief is good enough to send, but only because I absorbed the inconsistency. What I want is simple: people can keep using AI, but every draft should arrive with the same source pack, data boundary, brand example, structure, and final check so I am reviewing the thinking instead of repairing the process.