AI for Revenue Recognition (ASC 606) Quiz

Answers and explanations appear as you take the quiz.

1. A $100,000 contract bundles a license with a standalone selling price of $80,000 and support with a standalone selling price of $40,000. About how much of the price is allocated to the license?
   • $80,000
   • $50,000
   • $66,667
   • $100,000
2. Under ASC 606, a promised good or service is distinct when it meets which two conditions?
   • It is capable of being distinct and distinct in the context of the contract
   • It is invoiced separately and delivered within 12 months
   • It has a list price and a separate SKU
   • It is approved by the auditor and the customer
3. Why does the guide recommend running transaction price allocation in deterministic code or the revenue subledger, not in the language model?
   • Language models cannot read pricing tables
   • ASC 606 prohibits software from doing calculations
   • Allocation is optional under ASC 606
   • Model arithmetic can be wrong, and deterministic calculations are repeatable and auditable
4. Which limitation of AI contract extraction does the guide highlight as especially likely to change the accounting?
   • It cannot read PDFs
   • It cannot find side letters or informal concessions that were never put in the contract repository
   • It only works for SaaS contracts
   • It always overstates transaction prices
5. A contract gives the customer an option to renew at a steep discount not offered to similar customers. What should the AI flag it as for accountant review?
   • A possible material right
   • An assurance-type warranty
   • A lease component under ASC 842
   • A contract that fails step one
6. When is variable consideration included in the transaction price under ASC 606?
   • Always at its maximum possible amount
   • Only after cash is collected
   • Only to the extent it is probable that a significant reversal of recognized revenue will not occur
   • Never, because variable amounts are always excluded
7. Which design feature most directly lets a reviewer or auditor check an AI-extracted contract term?
   • A higher model temperature setting
   • Storing only the final revenue schedule
   • Letting the model rewrite the contract in plain English
   • A citation linking each extracted value to the document, page and text span it came from
8. What completeness check does the guide suggest for an AI contract extraction process?
   • Reconciling extracted total contract value to billing and CRM records
   • Asking the model whether it missed anything
   • Counting the pages in each contract
   • Comparing results with last year's revenue growth rate