what should I automate first
The first workflow to improve is rarely the most impressive one. It is the one where people repeatedly gather the same information, apply reasonably clear rules, and still have a human who can judge the result.
( The Detail )
Choose friction with a clear owner
That is why reporting hand-offs, supplier-cost checks, product-information clean-up, lead triage, and internal knowledge retrieval are often better first projects than a customer-facing agent with broad authority.
Use four practical tests
Assess each candidate workflow on business value, information readiness, risk, and ownership. A high-value process with no reliable source data is not ready. A low-risk process with no meaningful outcome is not worth building. The strongest first project balances all four.
If the team cannot explain what a good outcome looks like, or nobody is accountable for acting on the result, start with a workflow map instead of automation.
Keep the first loop visible
Build a narrow loop that a real owner can inspect: source information in, a clear output, human review, and a measured decision. Once that works, the same knowledge and operating pattern can support more ambitious systems.
( Next Step )
Start small enough to review, but on a workflow important enough to show whether a better system is worth building.