Client calls for AI automation projects often go sideways because

🔥

Client calls for AI automation projects often go sideways because both sides talk past each other. The client wants outcome 'less manual data entry,' the builder hears 'integrate X with Y.' A simple, memorable framework to keep the conversation focused on value is CLOSER, and you can run it without ever sounding slick. It's: Clarify why they're talking to you; Label their core problem back to them in their own words; Overview their past attempts and what specifically broke; Sell the outcome, not the tech stack (the vacation, not the flight); Explain only the concerns they raise, then ask for commitment; Reinforce the decision in the first 24 hours after they say yes. In automation, the Overview step is gold. Ask 'What have you tried so far?' and you'll hear about the Make scenario that double-fires, the Python script that works but only when Jim runs it, the vendor tool that lied about webhooks. That's your exact delivery spec, handed to you by the client. The Sell step then becomes almost unnecessary: reflect back the three problems they just told you about, connect each to a concrete fix (not the whole architecture), and they'll agree you understand it. The rest is just confirming budget and timeline. I've seen this turn messy discovery calls into 20-minute conversations that actually convert, without any pressure. The framework isn't about manipulation; it's about being so clear on their situation that they feel stupid hiring anyone else. Just state the facts as you hear them. Which part of client discovery do you find hardest to navigate without accidentally over-complicating the conversation?

2 comments

Client calls for AI automation projects often go sideways because