Deal Review Process: Catch Stalled Deals 3 Weeks Earlier

If you're a CEO sitting through a 90-minute deal review where the only output is "let's keep an eye on it," you already know the format is broken. The problem isn't the reps. It's that the meeting was designed to make everyone feel covered, not to surface stalls early enough to do anything about them.

Why do most deal reviews catch stalls too late?

Because the standard format invites narrative, not evidence. The rep tells a story about the deal, the manager asks a few clarifying questions, the room moves on. By the time the slip is undeniable, the close date has already moved twice and the champion has gone silent.

Definition: Deal stall — a deal that has stopped progressing through buying steps despite remaining in the same forecast category, usually visible 2-4 weeks before the close-date slip becomes official.

The signal is almost always there 3 weeks earlier — in the activity data, in the calendar gaps, in the email-thread length. The review just didn't ask the questions that would have surfaced it.

The 5-question deal review format

Every deal over the threshold gets the same five questions, no exceptions. The discipline of identical questions is what makes the review fast and the data comparable across deals.

Question 1: What is the buyer's documented next step?

Not the rep's next step. The buyer's. If the answer is anything softer than "the champion has scheduled X by Y date with Z stakeholders," the deal has a known stall risk.

Question 2: Who is the economic buyer and when did they last engage?

Name, role, date of last activity. If the rep can't answer all three within 10 seconds, that's a hygiene flag. If "last engaged" is more than 21 days ago, the deal is at high stall risk regardless of stage.

Question 3: What is the buyer's compelling event and how confident are we in the date?

Compelling event = the reason this has to close by a date, not "they want to start in Q4." If there's no compelling event, the close date is wishful.

Definition: Compelling event — a dated, externally driven reason why the buyer must act by a specific time, e.g. a contract expiry, regulatory deadline, board meeting, or scheduled go-live the buyer cannot reschedule cheaply.

Question 4: What single risk could kill this deal in the next two weeks?

Forces the rep to name one falsifiable risk, not "execution risk" or "competitive pressure." If the answer is vague, the rep hasn't disqualified anything; everything is theoretically possible.

Question 5: What does the rep need from the manager in the next 7 days?

The output. Concrete, dated, owned. If the rep can't name a specific ask, the review didn't produce action.

How does AI-assisted prep change this?

The five questions become useful only if the manager walks in already knowing the activity ground truth. AI prep handles that 24 hours ahead.

The night before the review, an AI job runs against each in-scope deal and produces a half-page brief per deal:

Deal: [ACCOUNT — DEAL NAME]
Value: $[X]  Stage: [stage]  Days in stage: [N]
Buyer activity (last 30 days):
- [contact role] — [N emails sent, N replies, N meetings]
- [contact role] — [N emails sent, N replies, N meetings]
Email-thread reply latency (last 60 days): [median in days, trend]
Last meeting: [date], [duration], [attendees]
Documents shared: [list]
AI-flagged risks:
- [Specific concrete risk, e.g. "no contact with finance since [date]"]
- [Specific concrete risk]
AI-flagged stall signal: [strong/moderate/none], reason: [text]
Rep's CRM next step: [text] (last edited [N] days ago)

The manager reads the brief in 90 seconds before the meeting and walks in with the falsification angles ready. The rep walks in knowing what's in the brief — no surprises, but no escape from the data either.

Tool tip (AIAdvisoryBoard.me): This is exactly what the Plan → Fact → Gap loop produces inside the daily-management OS. Plan is the rep's CRM record. Fact is the AI-assembled activity reality. Gap is what the manager walks in already knowing about. The deal review stops being a storytelling session and starts being a falsification session — five questions per deal, six minutes per deal, thirty-minute meeting wrapped. See how the 7-day diagnostic surfaces the same loop across every team at https://aiadvisoryboard.me/?lang=en.

Good vs bad answers

Question 1 (buyer's documented next step)

• Bad: "I'm going to follow up with Sarah this week."
• Good: "Sarah confirmed Tuesday's call with their CFO; agenda is signed, deck is reviewed."

Question 4 (single risk)

• Bad: "There's some procurement risk."
• Good: "Procurement won't engage until legal finishes their review of the MSA, which is sitting with their outside counsel and we have no visibility on her timeline."

The pattern: bad answers are unfalsifiable, good answers are specific enough to be wrong about.

Manager scan (2-minute digest example)

• Plan: 14 deals in this week's review, total $1.8M, all marked "on track" by reps
• Fact: AI prep flagged 6 deals with stall signals — 2 strong, 4 moderate
• Gap: Reviews must spend 70% of time on the 6 flagged deals; 8 healthy deals get a 2-minute confirm each
• Plan: 4 reps each said "champion is fully engaged" on their flagged deals
• Fact: Activity logs show champion last replied 18-31 days ago in 3 of 4 cases
• Gap: Champion definition needs a refresh — "engaged" = replied in last 14 days, not "knows me"
• Plan: All 6 flagged deals stay in commit
• Fact: None of the 6 have a documented compelling event
• Gap: Moving 4 to best-case this week, 2 stay in commit pending compelling-event check by Friday
• Action threshold: any deal that survives 2 reviews with the same flag goes to deal-clinic

Micro-case (what changes after 7-14 days)

A 95-person managed-services firm ran the standard 90-minute Monday review and consistently saw close dates slip 3-4 weeks after the review where the deal was last marked healthy. Sales leadership rewrote the format to the five-question structure and added a small AI-prep job that ran Sunday night. Week 1 review: 40% of in-scope deals got flagged with stall signals nobody had named out loud before. By week 3 the meeting ran 32 minutes instead of 92 and the reps started self-correcting next-step language before the review because they knew the AI brief was coming. By week 8 the close-date slip rate had moved from 31% to 14% on deals above the threshold, and two deals that previously would have slid to next quarter were closed in-quarter because the compelling-event question forced a real conversation eight weeks before close instead of three weeks before.

Note on this case: This example is illustrative — based on typical patterns we observe with companies of 30-500 employees, not a single named client. Specific numbers are rounded approximations of common ranges, not guarantees.

Tool tip (AIAdvisoryBoard.me): Many sales leaders ask whether AI replaces the manager's intuition in deal reviews. It doesn't. It replaces the 40 minutes the manager used to spend pulling activity data from the CRM, reading email threads, and checking calendars before the meeting. That work moves to a machine; the manager's time moves to the falsification questions only humans can ask. Plan → Fact → Gap is the framing that keeps the human in the loop where they add value. See it at https://aiadvisoryboard.me/?lang=en.

FAQ

Why exactly five questions and not seven or ten? Five fits in six minutes per deal with no rush, which makes a 30-minute review of 5 flagged deals realistic. More questions invite story; fewer miss the compelling-event check. The five chosen here are the falsification-heaviest of the standard MEDDIC/MEDDPICC sets.

What if a deal is too early-stage for these questions? Use a different review for early-stage. The five-question format is calibrated for deals in mid-funnel and later. Early-stage deals need a "did we qualify?" review, not a "will it close?" review.

Won't reps push back on the AI prep as Big Brother? They push back the first two weeks. They stop pushing back when they realize the brief gives them ammunition for the deals where they're right — and saves them from being caught flat-footed on the deals where they're wrong. The framing matters: brief goes to the rep at the same time it goes to the manager.

How does this fit with our existing forecast call? Forecast call asks "are these numbers reliable?" Deal review asks "what do we do about specific deals?" Different cadence, different audience, different output. Don't merge them.

What if our AI tooling is basic — can we still do this? Yes. The five questions work with no AI. The AI prep is what compresses prep time from 40 minutes to 90 seconds; the questions themselves are the structural fix.

Conclusion

Deal reviews don't fail because reps don't know the deals. They fail because the format doesn't force falsifiable answers and the manager doesn't walk in with the activity ground truth. Five questions, AI prep the night before, 30 minutes max, focused only on flagged deals.

Pick the five questions. Run the format Monday. Watch what the meeting produces in week 3.

If you want a system that surfaces the Plan → Fact → Gap automatically — every day, across the company, not just in deal reviews — see how the 7-day diagnostic works at https://aiadvisoryboard.me/?lang=en.