Coming Soon · Free
Capture the processes trapped in your head as procedures a cold-start AI agent can run — not docs only a human can read.
A working session for operators who want the highest-leverage processes out of their head and into procedures an AI agent can execute cold. Friday helps you rank what to capture by delegation pain — what an agent could take off your plate first — then captures each as a strict 12-field SOP, role-plays a zero-context agent to break it, and hardens it until it survives a cold-start run without stalling, guessing, or fabricating. You leave with an SOP Library: a prioritized delegation ledger, each SOP carrying the bottleneck it retires, a Break log proving it survives the run, and the verification step that confirms it worked. An SOP isn't done when its boxes are filled — it's done when an agent handed nothing but that document runs it end to end.
The arc
Calibrate the one signal that drives every downstream phase — who consumes these SOPs, an agent or a hire — and ingest a workflow-mapping matrix if you have one, or run a 10-minute inventory-lite if you don't.
Build the delegation ledger — a prioritized queue of what comes off your plate first, force-ranked on delegation pain, not importance. Each SOP earns its slot by naming the bottleneck it retires. Cap: 3 self-serve / 5 facilitated.
Populate the canonical 12-field schema for each selected SOP at the right granularity — the TWI Goldilocks band, anchored to one question: what does the agent need so it doesn't have to guess? Draw out the actual process, weird workaround included. Per-SOP, exploration mode.
Also from Framework Friday
What you'll walk away with
Score your workflows against five operator-tier criteria, map your top priority end-to-end, and produce a spec with governance that a builder could pick up tomorrow. Yours to keep, refine, and put to work — whichever way you choose to run it.
Friday role-plays a zero-context agent and tries to break each SOP — narrating exactly where it would stall, guess, or fabricate. Run per-step, not per-SOP. Fills the decision-rule and verification gaps, captures real failure history, and produces the Break log that proves cold-start survival.
Run the five binary screens out loud, per SOP, in place of a numeric grade — named owner, every decision ruled with a catch-all, a whole-SOP verification, one real failure mode, and a Break log with zero open breaks. The library is Agent-Ready only when every SOP clears all five.
Hand back the SOP Library as the delegation ledger — the first N things you hand to an agent, in order, each carrying its bottleneck and frees-up line — re-run the five-screen readiness out loud, and name the FridayOS bridge once as completion of started work.