Usage-based revenue recognition for AI-native companies
Metered pricing, resold model tokens, prepaid credits, and outcome-based agent fees collided with a standard written for fixed-price contracts. Here are the six judgment calls where AI-native finance teams actually spend their close — with the codification, the entries, and the fix for each.
The judgment calls that stall an AI-native close
Each links straight into the full analysis — theory, a realistic scenario, a decision tree, and the operational fix.
A big upfront fee to fine-tune a custom model — its own deliverable, or spread over the usage term?
Billing per resolved ticket or successful action — how do you recognize revenue you only earn if the agent succeeds?
Reselling third-party model tokens and GPU compute with a markup — report the gross billing or just your margin?
Prepaid credit packs that expire — when do you recognize the credits customers never use?
Billing dates that don’t line up with month-end — do you recognize usage delivered but not yet invoiced?
Sales commissions on contracts with no committed term or volume — capitalize them, and over what period?
Every question, worked four ways
This isn’t a summary of the standard. Each question is worked end-to-end in the same four-part structure — so you move from “what does the codification say” to “what do I book Monday” without leaving the page.
What ASC 606 actually requires, cited to the subparagraph — the recognition principle behind the question, in plain language.
A realistic AI-native contract with real numbers — the exact fact pattern that makes the call ambiguous in the first place.
The if/then logic your auditor will walk — which facts flip the answer, and where the judgment actually lives.
The fix — illustrative journal entries, the true-up mechanics, and how to keep the position audit-ready every close.
Not every question is equally common — or equally costly
Each question plotted on financial impact and operational effort; bubble size shows roughly how many metered, AI-native companies actually run into it. Cutoff and gross-versus-net touch nearly everyone; outcome-based pricing is far rarer today but carries the highest financial stakes.
Prevalence figures are directional Maximor estimates for metered AI-native companies, not survey data — read them as relative frequency.
Cited to the subparagraph — not the section
Every reference points to the specific subparagraph that carries the conclusion, cross-checked against two Big-4 handbooks. Where a call is fact-dependent, the guide says so rather than overstate certainty.
Take the full guide with you
The complete playbook — every question with codification references, journal entries, decision trees, and the operational fix.
- All six questions, fully worked
- Illustrative journal entries and true-ups
- Cross-checked against 2 Big-4 handbooks
See how Audit-Ready Agents handle these six questions on your contracts
Book a 60-minute Finance Transformation Audit — we’ll map your usage-based contracts to the right ASC 606 treatment, live.
Product by Outcome
