Everstage says 57% first learn a deal is at risk from a pipeline review or a rep. Time ten deals.

On October 23, Siva Rajamani, Everstage’s CEO and co-founder, published an essay on what revenue operations does once the AI is already running. The figures are company-reported. Everstage says each one comes from its Revenue Execution 2026 survey. Its note on that survey describes 195 senior revenue leaders at North American SaaS and AI-native companies, fielded in 2026. The essay says Everstage partnered with Emporia Research to run it.

In that survey, all 195 already run AI somewhere in the revenue motion. Fifty-seven percent first learn that a deal is at risk from a pipeline review or from a rep.

What the survey counted

Everstage reports an average of 4.3 AI use cases at a company. RevOps runs 5.4. The sales team can see 3.3 of them.

Once a risk signal appears, 45% of the leaders take four or more days to act.

The revenue figure is an estimate those leaders gave Everstage. Three in four put the gap between knowing and acting at 5% of revenue or more. One in six put it above 15%. Everstage sets those shares against a company with $100 million of revenue and writes $5 million to $15 million a year. That multiplication is Everstage’s. The essay names no company that lost either sum.

Everstage reports that three in four RevOps leaders spend two full working days firefighting and keeping the revenue system running, and that three in four spend 16 or more hours a week on manual work.

RevOps rates its own risk detection 3.3 out of 5. Sales rates the same system 3.7 out of 5.

Which lines are the survey

The essay also tells a weekday story. A tool flags a strategic deal on Tuesday at 10:04 a.m., nothing changes, and a rep raises it in pipeline review the following Monday at 9:30 a.m. Later in the essay, Everstage draws four stages for a single alert and prints “Illustrative · not survey data” on the drawing. The stages run from a flag that fires, to a person who sees it, to a change in what the rep does, to a change in the outcome.

Use the 57% and the figure that 45% take four or more days as the survey. Use the Tuesday story and the four-stage drawing as Everstage’s picture of the same claim.

Rajamani argues for a job he calls Revenue Architect, scored on whether a design changed what the seller did in time to matter. The sheet below times that on your own deals.

Time ten deals

Take the last ten deals your team marked at risk, or that moved a close date. Give each deal four cells. Pull the dates from the CRM and from the notes of the last forecast call.

1. System. The earliest date a tool recorded the risk: a health field, an alert, or an automated note. If several tools write to the record, keep the earliest. Write “none” when you cannot find one.
2. Voice. The earliest date a person said it, in the review notes, a rep comment, or a message.
3. Next step. The date someone changed the deal after that. A meeting booked, a term changed, a stage moved, or an owner reassigned all count. Leave it blank when the record shows talk and no change.
4. Result. Yes or no. Did that change move the amount, the close date, or the forecast category?

The six-of-ten cutoff below is the rule for this sheet.

• Six or more “none” cells in the system column means the tools are missing deals you already worry about.
• A system date four or more days before the voice date means the alert sat on the record. Four or more days is the wait 45% of Everstage’s respondents already report. Your ten rows show whether these deals sit on that side of the line.
• A voice date with a blank next step, or a next step whose result is no, means the meeting named the risk and the deal stayed put.
• Ask one rep to open the alert RevOps trusts on a live account. If the rep cannot open it, you are looking at a local version of the gap Everstage reports between 5.4 use cases and the 3.3 a seller can see. Count your own screens before you repeat 5.4.

What to do this week

• Fill the ten rows before the next forecast call, and bring the sheet into the room.
• On one live account, ask the rep which risk alerts they can open. Write the list RevOps thought was already on that screen.
• For each red deal on the call, mark whether a tool or a person said it first.
• Keep at least 5% of revenue in the margin as the estimate three in four of Everstage’s respondents gave. If you write a dollar figure beside it, take that figure from these ten deals.
• Pick the row with the widest gap between the system date and the voice date. Name the person who will change the next step before the following call.

CRO of One sells a five-hour consult block at $400 an hour. A block can work toward written stage rules, a first forecast-risk review on your own pipeline, or one agent lane spec with its quality bar and handoff person. We agree which at the start, and a block does not guarantee forecast accuracy.