System
/
Memory
/
Forecasting

Forecasting

One definition of your forecast categories, the evidence each requires, and how a human override is reconciled with the system's view. Categories get interpreted privately, team by team, until someone has to reconcile the rollup. Your GTM system's memory holds the evidence bar for each category, applied continuously, with any disagreement made visible.

Used across
Learning from every outcome
One category carries most of the miss
Commit slipped 23% against a 6% target
Best case finished within 4% of plan
Overrides beat the system on one signal
Champion access overrides were right 8 in 10
The system had weighted it 3 times lower
Upgrade proposedv9v10
LEARNING
The memory upgrades
Memory
Forecasting
v10
Categories defined by the evidence they require
Champion access weighted higher in commitv10
Confidence stated alongside every call
Rep sentiment removed as a commit criterionv10
Override rules and conflict resolution defined
Approve upgrade to Forecasting?
The current way > AI added on > AI-native

The moments a forecast is called, and what backs it

Everyone agrees what commit means until you ask two managers to define it. The gap between those two answers is where most forecast pain actually lives.

01 | The Current Way

02 | AI Added On

03 | AI-Native

A rep sets a category

The same word, different bars

Commit means one thing on one team and something looser on another. The rollup adds them together anyway.

A probability instead

The model replaces 'Commit' with a probability, say 73%. It has no idea your team's rule is that nothing under 80% gets called Commit.

Category with stated evidence

Every territory uses the same evidence bar to call a deal 'Commit', so when the numbers roll up, they add together instead of stacking five different definitions of the same word.

The system and the rep disagree

Seniority settles it

A VP overrides the rep's gut call on a pipeline review, and nobody writes down whether the VP was right three months later.

A flag, not a conversation

The forecasting tool flags the deal red. It does not say the champion has gone quiet for three weeks, so the rep has nothing to actually go and fix.

The gap is made explicit

The rep says the deal is closing. The system says a quiet champion for three weeks is driving the risk score. One reply from the champion would close the gap.

A manager overrides the call

The override vanishes

A judgement call gets made and the reasoning stays in the manager's head.

No mechanism to record it

The prediction is regenerated next week as though the override never happened.

Overrides are expected

Overrides are expected, recorded with their reason and reconciled against outcomes, because leaders own the commitment.

The forecast call is prepared

Reconstructed the night before

Hours spent rebuilding why deals moved, from CRM changes and memory.

Summarised, not evidenced

A neat account of what changed. It still cannot say which deals failed a stated evidence requirement.

Movement is already explained

Assessment runs continuously against Deal State, so each move carries the evidence that triggered it and nothing needs rebuilding.

The quarter closes

The 80% bar never gets checked

Win rate gets reviewed every quarter. Whether 80% was ever the right cutoff for Commit never does.

A sharper model, same cutoff

The model gets more accurate at predicting deals. Nobody checks whether 80% is still the number that actually separates Commit from Best Case.

Results reset the cutoff

Closed and lost deals show whether 80% still separates Commit from Best Case. Sales leadership approves the new cutoff before it changes what counts as Commit.

Referenced by

Capabilities that reason with this memory

Sales

Pipeline Review

Active
Two deals slipped out of forecast this week
Flagged before the meeting, with what's driving each
One at-risk deal carries a third of the number
Champion's gone quiet; the buyer isn't multi-threaded
Three deals need a next step set today
Sitting idle past the pace of deals that close
+
Your focus list for the review, in priority order
The deals that need you, and what to do on each
Focus list ready before the meeting
SEE HOW IT WORKS

Sales

Deal Health

Active
Northwind now at risk of going to a competitor
Champion's gone quiet; a rival is in the room
Two deals worth pushing while you're ahead
Momentum's with you, so press the advantage
One deal to cut: no budget, no path to the buyer
Sitting in forecast, propping up a false number
+
Every open deal, scored and ranked for the review
Where to save, where to push, where to walk
Where to spend your time, before the review
SEE HOW IT WORKS

Sales

Daily Brief | Sales

Active
Kestrel went quiet overnight
Champion didn't reply, and a new exec joined the thread
Two next steps from yesterday never got sent
Sitting overdue, and the deals are cooling while they wait
Solstice needs a follow-up before your 10am
Send it now, or lose the thread going into the call
+
Your three deals that need you today, ranked
In order of what's at risk and what moves the number
Your day, prioritised, before you open your inbox
SEE HOW IT WORKS
SEE ALL CAPABILITIES
Used by

Skills that use this memory

No items found.
SEE ALL SKILLS
The Platform

One system that understands, decides, acts and learns.

Every GTM signal flows through an AI-native operating layer into a system that runs on the surfaces your team already uses.

Explore the GTM System →
GTM Data & Knowledge
CRM · Emails · Calls · Marketing · Product · Support · Documents · Research
AI-Native Operating Layer
Context · Memory · Skills · Agents · Decision Traces
AI-Native GTM System
Understand
ICP Fit
Deal & Account State
Stakeholders
Risk & Health
Decide
Targeting
Pipeline & Forecast
Campaign Planning
Renewals & Expansion
Act
Outreach & Follow-up
Meeting Prep
CRM Updates
Alerts & Escalation
Learn
Upgrade ICP
Upgrade Messaging
Upgrade Playbooks
Upgrade Forecasting
Surfaces
CRM · Slack · Teams · ChatGPT · Claude · MCP · API
Output
Briefings · Artifacts · Alerts · Recommendations · Approvals · Actions

SEE WHY REVOPS + MARKETING LEADERS CHOOSE REVENUE LABS

The GTM teams that learn fastest will win. Build yours a system that learns.

An advantage competitors cannot buy back: years of success and failure, codified.

FAQ

What buyers ask

If it's not here, we'll answer it live.

Talk to us
What is forecasting memory?

One approved definition of your forecast: the categories, the evidence each requires, the confidence rules, what must be submitted and when, and how overrides are handled. Every capability that touches the forecast reads that definition rather than each carrying its own interpretation.

How do you make forecast categories consistent?

By defining them in evidence rather than in adjectives. Commit stops being a feeling once the evidence bar states exactly what must be true to claim it, and the assessment applies that same test to every deal in every territory.

How can AI improve sales forecasting?

Predicting harder does not help. A close probability does not tell a manager which evidence is missing or what would change the call. The useful contribution is applying the evidence bar continuously to live deal evidence, showing where the system and the seller differ, and naming what would resolve it.

Should a forecast ever be overridden?

Yes, and the rules should say so explicitly. Leaders own the commitment, so an override is legitimate rather than a failure of the model. What matters is that it is recorded with its reasoning and reconciled against the outcome, so judgement becomes evidence too.

How is this different from sales process?

Sales process defines how a deal should progress and what each stage requires. Forecasting decides what you will call that deal and commit to, given where it actually is. A deal can sit correctly at a stage and still fail the evidence bar for the category a rep has chosen.