Blast radius
A queue of real decisions — an agent auto-refund, a discount to close a deal, an all-customers comms send — needs sorting into Delegate, Approve, or Keep human. You're scored not on your gut but on each decision's true blast radius: the dollars at stake and whether it can be undone. The reveal is a delegation ledger — money exposed by over-delegating, hours lost by over-controlling. Runs entirely in your browser (0 uploads, works offline).
How to use it
- For each decision in the queue, choose who should make the call: Delegate (let the agent), Approve (human signs off), or Keep human.
- Hit Score my calls — each is checked against its true lane, set by real blast radius.
- Read the ledger: dollars exposed where you over-delegated, bottleneck hours where you over-controlled, and the risk you correctly avoided.
What this clears up (the fundamentals)
- Delegation is a risk decision, not a trust decision — the right lane depends on blast radius (dollars at stake) and reversibility, not on how much you "trust" the agent or person.
- The cost runs both ways — over-delegating a high-blast, irreversible call risks real money; over-controlling a low-stakes, reversible one turns you into the bottleneck. A ledger makes both visible.
- Reversibility changes everything — a cheap, undoable action is safe to delegate even if it's frequent; an expensive, irreversible one deserves a human even if it's rare.
- Gut feel is miscalibrated — people systematically keep low-stakes calls (control theatre) and wave through the occasional high-stakes one. Scoring against blast radius corrects both.
Where it's used
A leadership and AI-operations drill — the fastest way to build the instinct for what an agent should and shouldn't be allowed to do autonomously. It's a miatz build-lab concept playable; learn to build it yourself.
FAQ
What is "blast radius"?
The size of the consequence if a decision goes wrong — here, the dollars at risk plus whether the action can be reversed. Bigger, less-reversible consequences belong in more cautious lanes.
How is the score calculated?
Each decision has a correct lane. Choosing a less cautious lane than correct is "over-delegated" and adds that decision's dollars to your exposure; choosing a more cautious lane is "over-controlled" and adds bottleneck hours. Matching the lane is a clean call.
Why penalise over-controlling at all?
Because keeping every low-stakes, reversible decision on a human's desk has a real cost: slower operations and wasted senior time. Good delegation frees humans for the calls that actually need them.
Is anything uploaded?
No. The queue, your choices and the ledger all run in your browser — nothing is transmitted or stored. Turn off your Wi-Fi and it still works.
Limits
A teaching model with fixed blast values and a single correct lane per decision — real calls have context, escalation paths, and org-specific risk tolerance. Use it to build the "delegate by blast radius" instinct, then tune the thresholds to your own org.
Related
Part of the Demystify Playgrounds. Explore the rest from the Playgrounds home.
Bookmark this page (Ctrl+D, or ⌘D on Mac) — it works offline the next time you need it.