AI · Automations

Playbook automations

Scheduled AI workflows · who runs what, on what cadence, last result

Total automations

14

across composer, intel, forecast & workflow

Healthy

13

on-track · running on schedule

Failing

1

at-risk · needs operator attention

Daily LLM calls

2,418

rolling 24h across all automations

Heads-up — automation needs attention
1 failing · operator decides next step
All automations
14 total · 13 healthy · 1 failing
Today’s schedule

Run times across the next 24 hours · each dot is one scheduled invocation

all times America/New_York
Morning brief generator
1 runs
Vendor relationship intel
14 runs
Risk synthesis
14 runs
Anomaly detection (charges)
17 runs
Visual catalog match
1 runs
Approval router
9 runs
Retainer health forecast
1 runs
00
 
 
03
 
 
06
 
 
09
 
 
12
 
 
15
 
 
18
 
 
21
 
 
Healthy runFailing run·On-event automations (e.g. contract scans) only appear when triggered
Cumulative effort saved

Estimated producer hours reclaimed by automation this month

Total hours saved

~412h

across all automations · vs producer baseline

Brief intakes auto-categorized

8

categorized · approved by producer

Anomalies surfaced

47

flagged for human review (vs 0 missed)

Recap decks pre-drafted

3

scaffolded · producer finalizes copy

See also

Each automation produces flags — see the human action taken on every one in the decision log.

AI decision log

AI flags, doesn’t fix. Every automation surfaces — operators decide.