Ticket Triage Bot

IT · Marcus Bell · Classify inbound service-desk tickets, set priority and route to the right queue before a human reads them.

SampleProvenExpand

1 open issue

Running at 3.5x the registered volume. The registered claim understates its value.14 Oct 2026

Claimed

6 min

of a person’s time per run, 400 runs a month.
$2,480/mo if true.

What happenedProven

5.8 min

the work is actually worth

97%

62 of 1,372

“successes” its own output denies

4.5%

51 outputs are worth 5.8 min of a person's time, against a 6 min baseline — 97% of it. The claim holds up.

Cost

$182

a month. $0.13 per successful run.
Worth $8,223.

Recommendation

Expand

$0.13 per successful run against $6.20 of human time, 97% success, and 1,413 runs against 400 claimed.

Every measurement

One dot per priced output, last 30 days. The median is what the money follows — the agent's own 35s per run is machine time, not saving.

One dot for each of the 51 outputs we measured. Darker areas are where many landed on the same value.

Cost per run · 90 days

90 days ago $0.13now $0.13

Tools

3 approved · 97% success · last run today

jira_search1,413
jira_write1,413
confluence_read577

▶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.

5.4mCategory plus severity. Quick but real.25 Oct 2026
5.9mPriority, queue and a one-line reason.25 Oct 2026
6.5mCategory plus severity. Quick but real.24 Oct 2026
4.6mCategory plus severity. Quick but real.24 Oct 2026
4.7mRouted with a short justification.23 Oct 2026
5.3mPriority, queue and a one-line reason.23 Oct 2026

▶Recent runs

When	Outcome	Cost	Measured	Output
25 Oct, 20:54	success	$0.14	—	not captured
25 Oct, 20:30	success	$0.12	—	not captured
25 Oct, 19:17	success	$0.13	—	not captured
25 Oct, 19:06	success	$0.12	—	not captured
25 Oct, 18:48	success	$0.13	—	not captured
25 Oct, 17:53	success	$0.14	—	not captured
25 Oct, 17:51	success	$0.14	—	not captured
25 Oct, 17:34	success	$0.13	—	not captured

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