Why writing your own prompts doesn't make sense

May 6, 2025

I’ve been seeing lots of AI outbound tools that put the burden back on the user...

...asking them to build their own prompts as part of a fancy new workflow.

Here’s why I don’t think it makes sense:

1. Models evolve, so prompts need to as well.

No one in Demand Generation or RevOps has time to constantly rewrite prompts every time OpenAI releases a new version.

2. Users aren’t taught to prompt “the right way.

”There’s no standardization across products, and users shouldn’t have to be prompting in the first place.

3. Allowing users to prompt also means unlimited untested cases.

When you leave prompt creation wide open, you’re also opening the door to fake news, awkward wording, or things that sound too “AI”.

For an AI outbound solution to work day 1 to day 365, it has to be usable, maintainable, and drive real results.

For Rocketbrew, this means:

  • Users update a knowledge base of sales and marketing content instead of endless prompts
  • Users upload real outbound messages that worked, and our model generates lookalike versions
  • Users get a dedicated CSM reviewing monthly results and recommending updates

No prompt engineering degree required!