"We need an AI strategy" is usually the wrong first sentence
"We need an AI strategy" is usually the wrong first sentence, and it produces a predictable document: a list of AI capabilities with your company's name typed into the gaps.
Start somewhere else.
List the five most expensive workflows in the business. Not the most annoying. The ones where a person's time, a delay, or an error costs the most per occurrence. Claims review. Quote turnaround. Onboarding a new account. Reconciling the month.
For each one, write down where it waits. Waiting for a document. Waiting for an approval. Waiting for someone to re-type something from one system into another.
Now write down which decisions inside those workflows are made the same way every time by someone who could describe the rule.
That list is your AI strategy. It is short, it has dollar values next to it, and it will tell you quickly that two of the five items need a process fix and no software, one needs a $30 a month tool, and one might justify a real project.
An AI strategy that starts from capabilities finds places to put AI. One that starts from cost finds places where money is leaking, and then asks whether AI is the cheapest patch. Those produce different budgets, and only the second one produces a number the CFO will recognise a year later.