Before you take a proposal into a leadership meeting, give an AI reviewer the draft, the numbers behind every claim, and the roles in the room, and ask it to list every claim that would fail under their questions. This guide covers an executive staff meeting, a budget ask to a vice president, or a product review with the people who can stop the plan. For a formal board pack, see reviewing a board pack before the meeting. For a written reply after a board meeting, see reviewing your reply to a board question.
Where leadership proposals fail in the room
Leadership meetings are short. A proposal often dies on one slide that states a number the backup file doesn't support, one alternative the room already prefers that the deck never names, or a risk with no owner and no date for the next check. The person presenting usually wrote the deck with teammates who already agree, so the soft spots stay invisible until someone outside that group asks the first hard question.
A worked example
This example is illustrative and doesn't describe a customer. A product director has a six-slide proposal for Monday's executive staff meeting: add a paid support tier at $200 per month for mid-market accounts that already open more than twenty tickets a quarter. Before the meeting, the director puts five things into one request:
- The six slides and the one-page pricing sheet.
- An export of support ticket volume by plan for the last two quarters.
- Churn by plan from the billing system for the same period.
- The ask in one sentence: approve a three-month pilot with twenty accounts and a go or stop review after the pilot.
- One instruction: read as the chief financial officer and the head of customer success, list every claim the attachments don't support, and list the first five questions each of them would ask.
A review that helps points at specific lines. It might find that slide 3 claims mid-market accounts with paid support renew at 12 points higher than the rest of mid-market, while the churn export only covers two quarters and mixes mid-market with enterprise. It might notice that the pricing sheet assumes every pilot account stays on the tier after month three, with no line for accounts that drop back to the free tier. It might also ask what happens to ticket load for the remaining free tier if the best accounts move into paid support, because the staffing plan on slide 5 only covers the paid queue. The director checks each point against the exports, replaces the 12-point claim with the actual two-quarter gap and a note that the sample is small, adds a line for accounts that leave the tier, and names the customer success lead as owner of the ticket-load check at week six.
A request you can copy
Below are a proposal I will present to leadership, the decision I am asking for, and the documents the proposal relies on. The people in the room are [roles, for example the chief financial officer, the head of customer success, and the chief executive]. Read the proposal as each of those people in turn. For each role, list every claim the attached documents don't support, quoting the slide or line. Then list the first five questions that person would ask, and for each one say whether the proposal already answers it and where. Name the status quo and any rival option the proposal rejects, and say what tradeoff the proposal gives for that rejection. Don't rewrite the slides.
Leave the source files in even when the slides already quote them. Objections often come from a gap between a slide and the backup, and a reviewer that only sees the slides can't find it.
How the Pingpong web review app runs it
The web review app sends your request through several models in order. Each later model receives the original request and every earlier answer, with instructions to assess the work so far. With a proposal, an early pass might list unsupported claims by role, and a later pass might check whether a flagged number is wrong or only defined differently across files. You get a final answer and can open each earlier pass. Each reviewer reads the earlier answers, so a confident mistake in the first pass can carry forward. Check every flagged figure against the source file before you change a slide.
What a model review can't check
A model doesn't know which person in the room has already decided against the idea, or what was said in a hallway conversation last week. It also can't tell you whether the ask belongs in this meeting or needs a longer review with finance or counsel first. Anything that touches a pending legal matter, unpublished results, or a personnel decision should go past the people responsible for those areas before the slides go out.
If the proposal is a choice between options that hasn't reached a meeting yet, see reviewing a decision between options. If the stakes are high and you want a fuller package check, see reviewing a high-stakes decision package. For a general check on any draft, see getting a second opinion on a draft. For the steps in the app, see running your first review. More guides live under work decisions before you commit.