← Home

AI Operations Tools

AI automation is capital allocation. The same math your CFO uses for factories, fleets, and production lines applies to knowledge work automation - with one twist: these assets can appreciate.

“What accuracy do we need?” is the wrong question. The right one is what errors cost, how that compares to alternatives, and which investments change the equation. Models depreciate through distribution shift and competitive catch-up; verifiers, data moats, and institutional knowledge compound. That dual curve is what makes AI capital allocation different from buying a machine.

Start here
0
Classify your lineWIP

What Factory Are You? →

The five steps below price a single task. This one zooms out and classifies your whole line - which physical factory it resembles, and which century-old operations playbook transfers. Run it first to know what kind of line you are optimizing.

Then price each task · 5-step sequence
  1. 1
    DIAGNOSE

    Is this automatable?

    The Verification Quadrant →

    Map any task by calculation difficulty vs. verification difficulty. The Templeton Ratio determines whether AI saves you time or doubles it.

  2. 2
    CALIBRATE

    What are the stakes?

    The Dollarized Confusion Matrix →

    Replace accuracy with dollars. Calculate optimal thresholds from the actual cost of being wrong in each direction.

  3. 3
    INVEST

    What moves improve your position?

    Quadrant Shifting →

    Five capital investments that shift tasks to better quadrants. Build verifiers, decompose tasks, enrich inputs.

  4. 4
    VALUE

    What is the spread?

    Automation NPV →

    Calculate the risk-adjusted return on construction cost. Assess each opportunity, then rank them to decide where the next dollar goes.

  5. 5
    SCREEN

    Should you automate this at all?

    TaskVector →

    Score it across 9 dimensions. If any single dimension is a landmine, it is a hard no regardless of the composite score.

Why these five

Every step above encodes a specific failure mode I've seen kill AI projects inside PE portfolio companies - the Verification Trap, symmetric threshold bias, investing in models when the moat is in verifiers. The sequence makes those failure modes visible before you ship, not after.