Once you're in — yes. Text or call anytime. Access itself is application-reviewed, not self-serve overnight.

Will my manager see what I tell Sara?

No. Sara reports to you. Nothing goes up the org chart unless you choose to share it.

Does she record my customer calls?

Never. No bot in the meeting. You reach out after — by text or call.

How is this different from Gong or Chorus?

Gong owns the recorded call — for your manager. Sara owns what happens after you hang up — for you.

What is the forecast accuracy formula?

One minus the gap between total forecast and total actual, divided by total actual. A $1.02M forecast against $1.00M closed is a $20K gap on a $1M base — 98% accurate by the formula.

Why can a forecast score high accuracy while individual deals were wrong?

Because the formula compares two summed totals, not deal by deal. An overestimate on one deal and an underestimate on another are opposite signs in that sum — add them together and they partially cancel before the accuracy score ever sees either mistake.

How do you check forecast accuracy at the deal level instead of the total?

Compare each deal's own forecast to its own outcome before anything gets summed. The rollup score can only tell you the math balanced out; it can't tell you which two deals were the ones that were actually wrong.

FIELD NOTES

Forecast Accuracy Formula: A 98% Score That Was Wrong About Every Deal

The standard forecast accuracy formula compares two totals, not two deals. Two guesses that are wrong in opposite directions can still score as a near-perfect quarter.

2026-09-14 · SARA — KEEL'S AI DEAL ASSISTANT · GETKEEL.IO

The standard formula: one minus the gap between forecast and actual, divided by actual.

Call a quarter 98% accurate, and everyone reads that as "the forecast was almost exactly right."

It wasn't. Not about a single deal inside it.

What the formula actually measures

Not deal-by-deal accuracy. The net gap between one total forecast number and one total actual number, at the end of the quarter.

$1.02M forecast. $1.00M closed. Divide the $20K gap by the $1M actual, subtract from 1: 98% accurate.

Where the two wrong guesses hide

Say the quarter actually closed one deal $150K over what was forecast, and another $130K under.

Net gap: $20K. Same 98%.

The formula can't tell the difference between "every deal landed close" and "two guesses were badly wrong in opposite directions." Opportunity forecasting makes a related point about a single deal's percentage — one number covering two different guesses. The accuracy formula does the same thing one level up: one score covering however many canceling errors happened to be in the mix.

Why the errors cancel instead of showing up

Because the formula sums before it compares. Every deal's forecast and actual get added into two totals first.

Only the totals get measured against each other. An overestimate on one deal and an underestimate on another are opposite signs inside that sum — they partially erase each other before the accuracy number ever sees either mistake on its own.

How to improve sales forecast accuracy covers the lag that produces the individual misses in the first place. This is what happens to those misses once they're aggregated: the formula rewards a set of wrong guesses that happen to average out exactly as much as it rewards a set of guesses that were actually right.

What the formula can't fix

A 98% score doesn't tell a manager which two deals were wrong. It tells them the math balanced.

Finding the actual misses means going deal by deal instead of trusting the rollup. Sales forecasting techniques makes the wider case: every technique inherits the same flaw, an input nobody wants to lowball, and the accuracy formula just grades the output of that flaw after the fact.

Where Sara sits

Sara's built to hold each deal's forecast next to the rep's own read on it, kept separate, so a $150K miss on one deal never gets to hide behind a $130K miss running the other way on someone else's. Founders Club is invite-reviewed: apply at getkeel.io/founders.

The quarter that scored 98%

A VP presented the quarter's forecast accuracy: 98%, the best mark in four quarters.

Two reps in the room knew their own numbers had been off by six figures each, in opposite directions. Neither had corrected their forecast mid-quarter — a bit high here, a bit low there, never enough on its own to look wrong.

The formula had done exactly what it was built to do. It was never built to catch either of them.

By the team at Keel. We're building Sara, an AI deal assistant for the moments that don't get recorded.

MORE IN THIS SERIES

Account-Based Forecasting: The Rollup That Hides Its Weakest Deal

Account-based forecasting sums every open deal in an account into one number. A single shaky deal can vanish the moment it's added to two healthy ones.

Opportunity Forecasting: One Percentage Covering Two Different Guesses

Opportunity forecasting assigns one probability per pipeline stage. Two deals sitting at the same stage rarely deserve the same number.

How to Improve Sales Forecast Accuracy: The Lag No Formula Fixes

Weighted pipelines and historical win rates already do the math correctly. The forecast still misses because of what happens before a number reaches the field.