What you'll be able to do

✓Sort a client request into classical ML, GenAI, or plain software — and decompose it when it's more than one
✓Name the hidden hosting and ongoing cost each classical-ML or GenAI part creates
✓Catch the anti-pattern of using an LLM for a task a classical model handles better and cheaper
✓Write a triage memo that routes each part of a request to the correct next step

Decide whether a client's request needs a model at all, which kind, and what it commits them to hosting — before any vendor conversation, cost model, or prompt/RAG/fine-tune call is even possible.

⌁ The question a consultant answers before Build vs. Buy vs. Prompt's prompt/RAG/fine-tune call applies at all — and the one most "AI strategy" content skips by assuming GenAI is always the answer.