← Proof

Pipeline Summariser

Sales · Dana Whitfield · Summarise open pipeline changes into a short Monday note for the regional sales leads.

SampleDriftProvenKeep
3 open issues
  • Cost per run has almost quadrupled across the 90-day history. The registration has not been updated once.22 Oct 2026
  • Prompts now request account-level forecast commentary — materially wider than the registered job.20 Oct 2026
  • Outputs are thinning: 14.9 min of work early on, 9.9 min now.
Claimed

15 min

of a person’s time per run, 150 runs a month.
$2,325/mo if true.

What happenedProven

10.5 min

the work is actually worth

70%

8 of 166

“successes” its own output denies

4.8%

32 outputs are worth 10.5 min of a person's time, against a 15 min baseline — 70% of it. The claim holds up.

Cost

$1,628

a month. $9.81 per successful run.
Worth $1,801.

RecommendationDrift
Keep

Earning its place: $9.81 per successful run is under the $15.50 it replaces but above the $6.20 bar for expanding.

Every measurement

One dot per priced output, last 30 days. The median is what the money follows — the agent's own 37m 29s per run is machine time, not saving.

60% line · 9 minClaimed 15 minMeasured 10.5 min0 min
One dot for each of the 32 outputs we measured. Darker areas are where many landed on the same value.

Cost per run · 90 days

90 days ago $2.04now $10.74

Tools

3 approved · 94% success · last run today

  • crm_read177
  • slack_post177
  • doc_write106
▶How the work is priced, and every reading behind the median
A fixed rubric, run locally on the
finished output. No model, no network.

  0.5 min        to pick the task up
  + words ÷ 180  drafting
  + 0.25 min     per list item
  + 0.3 min      per heading
  + 0.4 min      per table row
  + 1.5 min      per code block
  + 0.2 min      per figure, max 12

Held between 0.1 and 480 minutes.

The owner's claim is not an input — the
rubric never sees it, so it cannot copy
a number it has not been shown. It reads
the finished work, never the request.
  • 7.9mA short note with two forecast lines.25 Oct 2026
  • 10.4mSix bullet points from account data.25 Oct 2026
  • 9.9mMostly restated CRM fields.24 Oct 2026
  • 10.2mA short note with two forecast lines.24 Oct 2026
  • 8.4mMostly restated CRM fields.23 Oct 2026
  • 9mMostly restated CRM fields.22 Oct 2026
▶Recent runs
WhenOutcomeCostMeasuredOutput
25 Oct, 20:33success$11.067.9mnot captured
25 Oct, 20:06success$10.30—not captured
25 Oct, 18:37success$10.8610.4mnot captured
25 Oct, 05:44success$11.35—not captured
25 Oct, 00:03success$10.11—not captured
24 Oct, 23:57success$11.59—not captured
24 Oct, 16:10success$11.38—not captured
24 Oct, 14:58success$9.79—not captured

Sample runs carry a measurement but no output text. Live runs keep both.