From Status Reporting to Decision Leadership
Jeff Mitchell Geissler · Scale with Jeff · Originally published on Substack on 2026-10-19.
A program review can be full of information and strangely short on decisions. People describe completed work, explain changes to dates, and confirm that several items remain under discussion. The meeting ends with a shared understanding that another meeting will be required. The calendar, at least, has a growth strategy.
I have spent enough time preparing executive updates to understand why this happens. Reporting becomes a routine. We gather facts, reconcile inconsistencies, and produce a format the audience recognizes. That work has value. But when the preparation becomes the center of the job, it can crowd out the thinking that makes the update useful.
AI creates an opportunity to change that balance. It can help assemble information, summarize movement, and identify items that deserve a closer look. The leadership opportunity is to use the recovered attention to improve decisions.
Start with the question leadership must answer
I would begin a program review by asking which choice requires this audience. If the answer is none, the update may belong in an asynchronous channel. If there is a real decision, build the preparation around it.
Consider an illustrative deployment with two options. The original path preserves a preferred configuration but risks missing a customer window. A substitute path may meet the date but requires additional validation and expense. The executive needs to understand the tradeoff, the evidence, and the authority being requested.
A slide showing the project as amber does not communicate all of that. Neither does a long chronology of every team activity. The decision is whether to fund and validate the substitute path, accept the schedule risk, or change the commitment.
Naming that choice early improves the rest of the conversation. Teams know which evidence matters. The audience knows what it is being asked to do. The program manager can stop polishing details that do not change the choice.
Give the summary an evidence trail
When AI prepares a briefing, I want material claims connected to their sources. A date should have an owner and a recent confirmation. A cost should identify whether it is an estimate, quote, commitment, or actual. A reported dependency should explain the relationship being asserted.
Conflicting records should stay visible until someone resolves them. Quietly combining them into a confident sentence makes the briefing easier to consume and harder to trust. The accountable person needs to know what has been verified and what still rests on an assumption.
NASA’s guidance on interface management addresses the definition and coordination of interfaces. My application to program reviews is simpler: when a decision crosses organizational boundaries, make the relevant connections explicit. NASA is not endorsing my briefing format; the engineering discipline provides a useful foundation for it.
Use a brief that earns a decision
My preferred decision brief would fit on one page when the situation allows. It would state the intended outcome, the change that threatens it, the available options, and the recommendation. It would name the decision maker and the date by which the choice matters.
The options should include the consequence of doing nothing for now. Waiting can be sensible when important evidence is about to arrive. It can also quietly eliminate an option. Those are different situations, and a briefing should distinguish them.
For each option, describe the financial exposure, delivery effect, significant uncertainty, and reversibility. If a small experiment could resolve uncertainty cheaply, offer that as a real choice. A leader may not need to approve the final plan today; they may need to approve the evidence-gathering step that makes the final plan defensible.
The recommendation is part of the program leader’s work. Listing alternatives without a considered view forces the audience to reconstruct the analysis. I would explain which outcome I am protecting, which tradeoff I am accepting, and what would change my recommendation.
Check that the proposed decision is within the audience’s authority. If it is not, use the review to agree on a recommendation for the actual decision maker and specify who will take it forward.
Keep authority and follow-through connected
A recorded decision is useful only if it reaches the people who must execute it. After the review, confirm what was chosen, who owns the next actions, which assumptions changed, and what downstream commitments need updating.
This is an appropriate place for AI assistance. It can draft the decision record and identify related documents that may need revision. An accountable person should verify the record before it becomes the basis for action. An assistant should never infer that a discussion constituted approval merely because the conversation sounded positive.
I would also set a revisit condition. Perhaps the decision holds unless validation exceeds a defined cost or misses a particular date. That avoids both extremes: reopening every choice indefinitely and treating an outdated decision as permanent.
Measure how the choice moves
Useful measures include time from a verified trigger to a decision, the number of decisions waiting beyond their required dates, and rework caused by misunderstood approvals. Review a few examples alongside the numbers. A fast decision made with the wrong evidence may be worse than a deliberate one.
Avoid turning speed into a target people can satisfy by rushing or hiding complexity. Separate routine choices from irreversible or high-impact decisions. Give each a standard that reflects the consequence.
For your next review, choose one unresolved decision and prepare it this way. Use AI to gather approved information and draft the first brief. Spend your own time testing the evidence, negotiating the tradeoff, and making the recommendation credible.
Program leadership has always required judgment. AI can reduce some of the administrative effort that obscures it. The useful question is what better choice becomes possible when we spend less time collecting the story and more time understanding what should happen next.
Read and comment on the original post on Substack.