Sztuczna inteligencja w quizie dotyczącym zakupów i analizy wydatków

Odpowiedzi i wyjaśnienia pojawiają się w miarę rozwiązywania quizu.

1. Jakie znaczenie ma restrukturyzacja podmiotu w analizie wydatków?
   • Assigns each invoice to a budget owner
   • Merges different spellings of the same supplier into one record
   • Predicts future prices for commodities
   • Checks whether invoices were paid on time
2. Co to jest „kostka wydatków”?
   • A view of spending by category, supplier and business unit
   • A secure storage format for contracts
   • A negotiation script used by chatbots
   • A legal requirement in public procurement
3. Jak według przewodnika najlepiej rozumieć ocenę ryzyka dostawcy?
   • A guaranteed prediction that a supplier will fail
   • A legal rating required by regulators
   • A prioritization tool that tells a team where to look first
   • A replacement for supplier contracts
4. Do jakiego rodzaju transakcji najlepiej nadają się agenci negocjacyjni AI?
   • Strategic, multi-year partnerships with key suppliers
   • High-volume, low-complexity deals with tail-spend suppliers
   • Mergers and acquisitions
   • Government defense contracts
5. Jakie limity zazwyczaj ustala kupujący dla agenta negocjacyjnego?
   • None; the agent chooses any terms it likes
   • Target terms, walk-away points and acceptable trade-offs
   • Only the supplier's name
   • The supplier's internal profit margin
6. Why is classification accuracy better measured by spend value than transaction count?
   • Transaction counts are not stored in ERP systems
   • A few large misclassified invoices matter more than many small ones
   • Regulators require value-based reporting
   • Spend value is easier for models to predict
7. In public procurement, AI is more often used for which purpose?
   • Automatically choosing contract winners
   • Analysis such as detecting collusion or fraud
   • Replacing public tender rules
   • Setting tax rates
8. Why do classification systems often stop at the deepest confident level of a hierarchy?
   • Deeper categories are illegal to use
   • A model may be sure a line is 'IT' but unsure whether it is software or hardware
   • Hierarchies have only one level
   • It makes processing slower on purpose