There are a few problems I see here...

- As others have pointed out, if you need to ask us how to GTM, then your product isn't effective, as your entire claim is that it will tell you.

- AIs hallucinate badly when it comes to numbers. The techniques for preventing hallucinations need an valid data set to contrast responses against, which either won't exist or exists because you have hard-coded valid GTM strategies into a data set. So the end result is either prone to hallucinations, or is just regurgitating a canned GTM playbook.

In short, while the idea is interesting, I can't imagine it working well in reality under the current LLM stacks.

Tks,

We didn’t hard-code GTM strategies, the agent uses strong prompting + internal logic to get high-quality output from the LLM. It’s not just regurgitating a playbook, but it’s also not freestyling.

Actually our GTM strategy does not cover the design of a high converting homepage. (it give the strategic alignement a.o) We could add that later and see the impact on our homepage.

Financial benchmarking use their specific models and fine tuning. We constrain them hard, structured prompts, known-good data, and refined knwoledge from the agentic system. No room for AI "vibes." The results and projection (financial strategy is not accounting) is actually quite impressive tbh.

For others agents we see hallucination has a strong feature, while obviously it's not the case for financial benchmarks.