Been using a couple different AI tools for copywriting work and keep running into the same friction. When I write for PT clients, the voice needs to be grounded, specific, anatomyaware. When I switch to B2B marketing content the same day, I need something punchier and more abstract. The tools I've tested don't shift registers cleanly between these two modes unless I do a lot of prompt massaging each time.
What I keep wondering is whether this is a prompting skill gap on my part, or whether current models genuinely flatten specialized professional voice into something generic. The output I get is competent, but it reads like it was written by someone who read about PT or B2B marketing rather than someone who actually worked in it.
There's also this weird thing where the more specific I get in the prompt, the more the model hedges and softens language that should be direct. Not sure if that's a safety tuning issue or just how these models handle professional domains.
Curious if others working across two pretty different professional contexts have found a tool or approach that handles this without needing a 400word system prompt every single session.
[link] [comments]