Which deals are about to slip, and how you see it in time
Omnia Lab is a sales control system that runs on top of your CRM — Bitrix24, amoCRM or any other. A deal is at risk when it has no next step: the customer got a reply but no task exists, a promise slipped, a missed call was never returned. The system catches it the same day.
· · Author: Dmitry Marenich
A deal rarely dies the day the customer says no. It dies earlier — when movement stops and nobody notices. Every evening at 18:00, Omnia Lab reviews the day's active deals on four axes: the rep, the customer, the deal, the process. Each deal gets a priority score from 0 to 100, and the report opens with one screen, "Needs attention today": up to ten deals scoring 70 or higher. The list is short on purpose. You should know where to start tomorrow morning, not where to start looking. And the risk does not stay a line in a report: for a deal on that list, the "Next step" recommendation lands in the CRM as a task with a mandatory deadline, and the rep gets it in the morning as part of the day plan.
What does the 6 PM deal review look at?
The first four items are the axes: they look at the same deal from different angles, so "stalled" never gets confused with "the customer is thinking it over". The fifth is not an axis — it is what the review produces.
- 01
The rep
How the conversation went: did they hear the customer, did they push for the next step, where did the tone slip. The rep score is 0.6 × conversation quality + 0.4 × process quality.
- 02
The customer
Who is on the other side and what they actually want. A customer profile is recorded with a confidence level and direct quotes: retail buyer, reseller, broker, or a buyer whose budget is nowhere near what they are asking about.
- 03
The deal
What it rests on: what was agreed, whether a deposit was taken, what is blocking the next step.
- 04
The process
Were the steps followed: was the reply on time, is there an open task, how long the deal has not moved.
- 05
What comes out
A report for the head of sales, personal tasks for reps, and a scoring history per deal. The card fields "Deal summary" and "AI analysis status" fill themselves, and existing comments are never overwritten.
How does the "Needs attention today" list get built?
This is the first screen of the report. It exists so nobody has to scroll the pipeline by eye.
- Deal priority
- 0–100, calculated during the evening review
- Threshold to make the list
- 70 and above
- List size
- 10 deals per day, maximum
- Rep score
- 0.6 × conversation + 0.4 × process
- Comment on the deal timeline
- when the rep axis falls below 50, or on a red flag
- Comment frequency
- no more than one per deal per day
- Night mode (optional)
- batch at 00:05, catch-up by 09:15, delivery at 10:00
In night mode the 18:00–21:00 window is still covered, and the AI cost is half.
What are the signs that a deal has stalled?
These signals do not wait for the evening. They fire during the day, and some of them close themselves once the rep takes a step.
- The customer got a reply, but there is no open task on the deal. The "No next step" block lands at 11:00 and 16:00.
- The customer went quiet. Follow-up nudges come on day 1, 3, 7 and 30, in one personal digest at 15:00 with buttons.
- The customer promised to come back and did not. A task "Customer promised to return" is created for their stated time plus 30 minutes. A vague "I'll think about it" creates nothing.
- A rep pushes their own deadline twice or more. That alone raises a separate signal in the "Important notifications" chat.
- A missed call was never returned. A "Call back" task is due in 30 minutes; at night Telegram stays quiet and the deadline moves to the start of the working day plus 30 minutes. A real conversation of 15 seconds or longer closes the task by itself — 5 to 8 seconds of ringing does not count.
- The call itself went badly. Follow-up calls generate a coaching task when the score drops below 65. First calls always get one.
How do you clear a deal off the list honestly?
Without a sign-off the list shows the same rows every morning. But a sign-off that never expires is just a way to hide a deal.
- 01
Five reasons to choose from
The head of sales marks the deal and picks one of five reasons. The reasons fall into two classes, and the class decides how long the mark holds.
- 02
Class A — handled
The head of sales has already stepped in. The mark holds until the first new risk appears; when it does, the deal is back on the list.
- 03
Class B — disagreeing with the system
The head of sales does not accept the assessment. The mark holds for as long as the risk doesn't get worse, and never longer than 14 days.
- 04
Reversible and on the record
The mark can be removed in one click. Who set it, when, and for what reason — all of it goes into the audit trail.
Which deals and reps does the system leave alone?
Half the trust in a review comes from what it refuses to touch. These exclusions are built in on purpose.
- Terminal and won stages. The daily review skips them — closures are handled separately, by close control.
- Reps flagged "do not remind". You decide who is on that list.
- Deals whose owner is not in the manager directory. Those raise no pings and never go into paid analysis.
- Pipelines outside the scope, and any stage on the configurable "ignore" list — no SLA, no digests.
- Any deal with a deposit. It is never rejected as junk and never moved to the cold pipeline.
- Failing to push an unqualified buyer is not counted against the rep. A mitigating segment only counts at 75% confidence or higher and with a direct quote: haggling, a "that's too expensive", short replies or silence never prove on their own that a buyer cannot pay.
What do the manager and the rep get after the review?
The head of sales gets the report, plus an "At risk" section in the panel. The panel reads a prepared data mart instead of querying the CRM live: refreshed hourly at :25, fully reconciled once a day, 90 days deep. The same evening, each rep gets a private "Day review" — one thing done well, one growth area with a specific angle. In the morning they get a day plan as a single message with four blocks: deals needing attention, promises due to customers today, tasks for the day, and recommendations with buttons. The "Next step" recommendation lands in the CRM as a task with a mandatory deadline and two fields: "Why now" and "Where to start". A new one closes the old one. Movement on the deal closes it by itself.
What changes in the numbers after rollout?
Measured before and after rollout at live customers. We do not name them.
- Median first reply to a customer
- from 32 minutes to 6.5
- Share of deals with a deposit
- from 4—7% to 13%
- Customer revenue
- +33% in 4 months, same headcount
- Rollout
- 3—5 days
Measured at live customers, among them a premium car dealership. Omnia Lab runs on top of the existing CRM rather than replacing it.
Частые вопросы
- Which deals count as at risk?
- The ones whose priority came out at 70 or higher in the evening review. That review looks at the deal on four axes: how the rep handled the conversation, who the customer is, what the deal rests on, and whether the process was followed. No more than ten deals make the list on any given day, so each one can realistically be dealt with.
- How do I tell a stalled deal from a slow one?
- A stalled deal is a deal with no next step. Twice a day, at 11:00 and 16:00, the system surfaces deals where the customer got a reply but no open task exists. On top of that it tracks broken promises, missed calls that were never returned, and customers who said they would come back and did not.
- Why only ten deals?
- Because nobody reads a hundred-row list. The 70 threshold and the cap of ten are set so the list can be worked through in one morning pass. Nothing disappears — everything else stays visible in the "At risk" section of the panel.
- What if the head of sales disagrees with the assessment?
- They mark the deal with a class B reason — disagreeing with the system. The deal drops off the list for as long as the risk doesn't get worse, and for no more than 14 days. The mark can be removed at any time, and every decision stays in the audit trail.
- The system writes into the deal card. Will it overwrite anything?
- No. It fills the "Deal summary" and "AI analysis status" fields, and comments are only appended, never rewritten. A timeline comment appears only when the rep axis drops below 50 or a red flag fires, and at most once a day per deal.
- Do we have to change CRM?
- No. Omnia Lab runs on top of the CRM you already use. Bitrix24 is the main integration, and amoCRM is on our integration list too. Stages, pipelines and the manager directory are configured for you, and all 33 modules have their own switches, so any function can be turned off.
Read next
- How to Train Sales Reps on Their Own DealsHow to train sales reps on their own deals: a private daily debrief, a four-block morning plan, and a Monday review where every claim carries a quote.
- Objective Sales Rep Scoring That Holds UpObjective sales rep evaluation: code computes the score, not the AI. The outcome sets the band, critical errors cap the score and need a verbatim quote.
- AI Sales Director: the system that runs every leadOmnia Lab is an AI sales OS that sits on top of your CRM. It routes leads, escalates at 5/10/15/17/20 minutes, tracks promises, and reviews closed deals
- Deals closed as lost with no reason: how to checkReps close deals as lost or spam to hide slow work. How to audit closed-lost reasons and put real deals back in play.
Let's see which of your deals are standing still
The free audit takes 3—5 days. We go through your team's real deals and conversations and show you where movement stopped, how many deals are in that state, and what to do about them. After that the decision is yours — rollout takes 3—5 days, and nobody will rush you.