When our department heads present operational bottlenecks during quarterly planning, how do we distinguish between an issue that requires fixing our Core Process and one that can actually be solved by an AI agent running a documented SOP?

Question

Tyler Smith · Accepted Answer

During your quarterly planning session or weekly Level 10 Meeting, department leaders often pitch software solutions for operational bottlenecks. If you apply automation to a broken process, you simply produce bad results faster. You must use IDS to get to the root cause before approving any automation project.

To determine whether you are facing a process failure or an automation opportunity, evaluate three criteria:

- Documentation status: Does the task follow a proven, documented standard operating procedure that already produces consistent results when executed by a person? If the answer is no, the process is flawed. You must define the process steps, clarify inputs and outputs, and test it manually before writing a single line of automation.

- Decision complexity: Does the workflow require subjective emotional intelligence, nuanced negotiation, or situational discretion? If yes, it must remain human. If the workflow consists of rules-based data validation, routine triage, or cross-system transcription, an AI agent can run it reliably.

- Financial impact: Never launch an AI project for the sake of novelty. Frame every initiative as an operations-improvement project. Calculate the exact labor hours freed on your P&L or the cycle-time reduction achieved.

If the process is already documented, strictly rules-based, and consumes expensive employee time on low-value coordination, assign a ninety-day Rock to deploy an AI agent. If the process produces inconsistent outcomes today, fix the standard operating procedure first.