March 12, 2026

Approval Gates: How Many Is Too Many

Every added approval step reduces risk and adds friction. The design question for any AI-assisted process is finding the smallest number of gates that still catches the mistakes that matter.

The case for fewer gates

A process with an approval step after every single sentence defeats the purpose of automating anything, the manager ends up doing as much work reviewing as they would have doing the task themselves. Over-gating a process is a common overcorrection after a single bad experience, and it usually kills adoption.

The case for more gates

A process with only one gate at the very end, after a long automated chain of research, drafting, and formatting, means a mistake introduced early can compound silently through several steps before anyone looks at it. By the time it surfaces, the fix requires redoing much of the chain.

A workable pattern

Two gates tend to cover most business-development workflows well: one after research and planning, before drafting begins, and one final approval before anything reaches a client or prospect. The first catches a wrong direction early and cheaply. The second is the non-negotiable check on anything client-facing. Anything beyond these two should be justified by a specific, recurring failure mode, not added by default.

Where teams get the gate count wrong most often

The most common mistake is not too many or too few gates in the abstract, it is placing the one unavoidable gate too late in the process, after a long chain of automated steps rather than right before the final send. A team can run an elaborate five-step automated research and drafting chain internally and still keep to a single external gate, as long as that gate sits at the true point of no return and nowhere earlier is treated as good enough to skip it.

A short audit worth running once a quarter

Pick a handful of recent client-facing deliverables and trace, in writing, exactly which gate each one passed through and when. If any deliverable's trail is unclear or the gate appears to have been skipped under deadline pressure, that is a more urgent finding than any theoretical debate about the right number of gates, and it deserves immediate attention regardless of how the rest of the process is designed.

A short note on the two-gate pattern in a very small team

A team of two or three people may find that the same person effectively sits at both gates, which is fine as long as the two checks happen at genuinely separate moments, planning review before drafting, final review before sending, rather than being collapsed into a single glance at the finished product.

A short note on communicating the approval framework to a new client

A prospective client asking how AI fits into a firm's process generally responds well to a short, specific answer describing exactly where the approval gates sit, rather than either an overly technical explanation or a vague reassurance. Being able to describe the process in two or three plain sentences is itself a sign the process is well designed, since an overcomplicated answer often reflects an overcomplicated, and less reliable, underlying process.

It is worth rehearsing that short answer before a client actually asks, rather than improvising it in the moment. A firm that has to think hard about how to explain its own review process has usually not thought hard enough about the process itself.

A final word on trust building over time

Trust in an agentic system should be earned incrementally and specifically, expanding what runs with less supervision only for the exact task categories that have built a real track record, rather than as a general, across-the-board loosening once a tool has performed well a few times. Specific, task-by-task trust is more resilient than blanket trust, because a failure in one category does not have to force a rethink of everything else the system does well.

VIPMarketing Editorial