A plan, not a prompt.
The seven steps I wish I'd had before my first AI agentic workflow.
Hey there,
I'm excited to share a new article this week, and it comes from a pattern I keep running into. A team gets excited about AI, opens a chat window and starts building, and a week later nobody can say how much to trust what they built. (I say "a team," but sometimes the team is me.)
So I wrote down the seven steps we've been using with partner churches to build AI agentic workflows, which are repeatable processes where AI handles a multi-step job on its own. The steps run from picking the right first task to deciding how much oversight the workflow needs, and you don't need a technical background to use them.
If you want a plan before you write the first promptSeven Steps to Build an Agentic WorkflowAI agentic workflows promise efficiency, but without a shared vetting process, they often create more friction than they solve. We built a framework for team...If you give it a try, which task would you start with? I'd love to hear.
Brad
P.S. I'm at CITN in Louisville Oct 4-6. Say hi if you'll be there.