RevOps playbook · AI Employee: Rev

Deal-at-Risk Detection

Save motions triggered earlier

Install this playbookBack to marketplace
The problem

By the time a deal is officially "at risk" — pushed a quarter, gone dark, lost to competition — it's too late to save. The signals that would have caught it two weeks earlier are all in the CRM and the engagement tools, but nobody looks at every deal every day. The manager finds out in the forecast call, when the rep says "pushed to next quarter" for the third week running.

At a glance
Trigger
Cron + CRM event
Approvals
None
What it does
Read-only
Systems
Salesforce · Gong · Product usage
How it feels in production

An hour-by-hour walkthrough.

Every morning at 07:00, Rev scans every open opportunity above $50K. For each, it runs a health check: - Days since last champion activity (email opened, call taken, meeting held) - Multi-threading depth (are we in the account beyond one contact) - Time in current stage vs. team benchmark - Close date slips in the last 30 days - Competitor mentions in Gong call transcripts - Sentiment shift on the last three touches A deal that crosses two thresholds flips to "at risk." Rev writes an annotated note on the opportunity, DMs the rep with the specific signals and a suggested next-best-action ("champion silent 12 days, competitor mentioned twice on last call — recommend exec sponsor outreach"), and cc's the manager if the deal is over $100K. The rep confirms or dismisses. Dismissals feed the model. Confirmations get a short recovery plan drafted for the rep to accept or edit: exec sponsor email, product-team involvement, discount review, or graceful stall (some deals should die faster). Weekly digest to the sales leader: at-risk value by segment, deals recovered, deals lost after flagging, and the top three signals doing most of the flagging work.
How it works

Step by step.

  1. 01

    Score every open opportunity daily

    Pull every open opportunity above the threshold. Run the health check: champion activity, multi-threading, time-in-stage, close-date slips, competitor mentions, sentiment.

    Salesforce · HubSpot · Gong · Outreach
  2. 02

    Cross-check against engagement + call intelligence

    Gong / Chorus transcripts give competitor mentions, sentiment, pricing pushback. Outreach / Salesloft gives cadence-response rates. LinkedIn Sales Navigator gives champion job-change signals — a champion leaving is often the earliest warning.

    Gong · Chorus · Outreach · LinkedIn Sales Navigator
  3. 03

    Flag + explain + suggest action

    Two thresholds crossed flips to at-risk. Rev writes an annotated note on the opportunity naming each signal, then DMs the rep with a suggested next-best-action grounded in what similar recovered deals did.

    CRM notes · Slack DM · Teams
  4. 04

    Draft recovery plan on confirmation

    Rep confirms → Rev drafts recovery plan: exec sponsor email (drafted, waiting for send), product involvement request (drafted, waiting for send to PM), discount analysis (numbers only, not sent). Rep edits and dispatches.

    CRM · Email · Slack · Approval flow
  5. 05

    Track outcomes + feed the model

    Every flag closes as: recovered, lost, dismissed-false-positive, or graceful-stall. Outcomes feed the model weekly; false positives tune thresholds; recovered patterns become suggested plays.

    Data warehouse · Model retrain
Systems and wiring

What you connect to make this run.

Salesforce · HubSpot

read+write

Read opportunity + activity + contact data. Write the health-check note + at-risk flag back so pipeline reports and dashboards see the same signal.

Gong · Chorus

read

Call transcripts + sentiment + competitor mentions. The single strongest signal is "customer said a competitor's name twice or more on the last two calls."

LinkedIn Sales Navigator · News APIs

read

Champion job-change alerts + company news (layoffs, funding, acquisition). External signals that the CRM never sees.

Slack · Teams · Email

write

Rep DM with signals + suggested action. Manager cc on high-value at-risks. Drafted recovery emails staged in the rep's outbox for review.

What changes

Before and after, honestly.

At-risk deals identified before slip
Before
20-40% (manager memory)
After
80%+ (daily scan)
Recovery rate on flagged deals
Before
10-20%
After
35-55% (earlier flag = more time to act)
Days from silence to intervention
Before
20-45 days
After
3-7 days
Forecast slippage per quarter
Before
20-40%
After
8-18%
Related

Playbooks that pair with this one.

RevOps

Weekly Forecast Roll-Up

Health signals here feed the forecast score there.

RevOps

Closed-Won Cascade

The healthy counterpart — what runs when the deal wins.

Frequently asked

Answers about this playbook.

Won't this cry-wolf on every deal that has a slow week?

Threshold is two independent signals, not one. A quiet week alone doesn't flag; a quiet week plus a competitor mention plus a close-date slip does. Weekly digest surfaces the false-positive rate and lets you tune.

How does it handle enterprise deals with slow, deliberate cycles?

The scoring model is stratified by ACV band. Enterprise deals get a longer baseline — 21 days of champion silence is normal for a 12-month cycle, not for a 30-day one.

Can reps opt out of certain deal types?

Yes — deals tagged as strategic / long-cycle / partner-led can be excluded from auto-DMs and only flagged in the manager's weekly review. Rep + manager set the policy per team.

What if the champion actually left the company?

LinkedIn signal fires the moment their profile updates. Rev drafts a "transition play": intro request to their replacement, farewell note to the champion, updated stakeholder map. Manager cc'd because these are high-mortality events.

Does this replace the manager's 1:1 with the rep?

No. It replaces the part of the 1:1 that's "walk me through your pipeline" — the answers are already in the dashboard. The 1:1 becomes about the plays: what did you try, what will you try next, what do you need from me.

See it run on your data.

Free plan, no credit card. Connect the systems this playbook needs and run it against a past event first.

Book a demoMore RevOps playbooks