Live · Free
What your AI actually costs — across model, infra, and ops time.
A self-serve diagnostic that surfaces your current AI cost across three layers — model spend, infrastructure, and the human ops time spent running, fixing, and supervising AI workflows. You walk through structured probing to inventory what you're actually paying for and where, and walk away with a one-page CFO-readable cost picture: total monthly footprint, 1-3 pattern badges, attribution coverage, locked-vs-flexible composition, and the cost-discipline questions to take into the next finance conversation.
The arc
A short read on you, the operating context, and what triggered this session. Calibrates probing depth for the rest of the session.
Inventory model spend + infrastructure together. Probes for buried items, shadow spend, and commitment structure (annual / monthly / usage-variable).
Human hours per week spent running, fixing, reviewing, iterating, onboarding others, managing vendors. The methodology's durable differentiator. Uses structured triangulation: round-number estimate → one time-study probe → revise if contradicted by >2× → capture genuinely-uncomputable hours as Unattributed Ops Time.
What you'll walk away with
Your real AI spend across model, infrastructure, and the ops time nobody counts — with the patterns and the questions to take into your next finance conversation. Yours to keep and rerun.
Also from Framework Friday
For top-2 named workflows: throughput → cost-per-unit (Pattern F). Plus the trajectory probe — what did this spend look like 3 months ago? Order-of-magnitude is fine.
Friday names the layer balance + 1-3 structural patterns + attribution coverage axis + commitment ratio + (Pattern E positive-criteria check if applicable). Patterns surface questions, not directives. CFO top-down artifact format.
Artifact download. Routing based on triggered patterns + operator profile. FridayOS bridge framed as durable cost visibility — same Friday, same methodology, persistent.