Task Logging → Automation Backlog

Type: Process pattern
Referenced from: audit-test-automate-ai-delegation

Definition

A systematic method for identifying automation opportunities:

  1. Log — For one month, write down every task you do, how long it takes, and everything you wanted to do but didn't
  2. Group — Dump all logs into an LLM, ask it to group them by category
  3. Analyze — Create a spreadsheet of internal processes and their time costs
  4. Test — Run each task through the 3-Questions Test
  5. Automate — Tasks that pass become the automation backlog

Why It Works

  • Exhaustive — Captures the full picture, not just what you remember
  • Data-driven — Time costs reveal where leverage exists
  • "Wanted to do but didn't" — These are the blocked tasks that automation could unlock
  • The log IS the backlog — No separate backlog creation step needed

Example from the Article

After one month of Google Keep logging, the founder discovered:
- 3 hours writing tests
- 2 hours reconciling bank accounts
- 10 hours reading
- 1 hour cleaning up tickets

This revealed which tasks had the most automation potential.

Application to dark-factory-kb

The KB pipeline follows this exact pattern:
- log.md = the task log (every operation recorded)
- reports/ = the analysis (lint reports, health checks)
- Concept extraction = identifying what's automatable
- Pipeline stages = the automation backlog (what's been automated)

The pipeline's append-only log IS the automation backlog — operations are logged, analyzed, and the pattern repeats.