Guide-linked quiz · Medium Level
Pricing AI Products: Seats, Usage and Credits Quiz
Tests why AI compute costs change software pricing, and the trade-offs of per-seat, usage-based, credit and hybrid pricing models for AI products.
Question 1 of 8
Why does AI make per-seat pricing riskier for vendors than in traditional software?
Preview every question8 questions
Answers and explanations appear as you take the quiz.
- Why does AI make per-seat pricing riskier for vendors than in traditional software?
- Seats are illegal for AI products
- Every AI request has a real compute cost, and heavy users can cost far more than light users
- AI users never log in more than once
- Traditional software has higher marginal costs
- What is the main drawback of usage-based pricing for buyers?
- It is less aligned with value
- It cannot be used for APIs
- It makes spending unpredictable and harder to budget
- It always costs more than seats
- What advantage do credits offer vendors?
- They let actions with very different compute costs be priced under one balance, with rates adjustable as costs change
- They remove all compute costs
- They guarantee customers understand exactly what they pay for
- They eliminate the need for metering
- What is the key customer complaint about credit systems described in the guide?
- Credits expire too slowly
- Credits are always more expensive than seats
- Credits cannot be bought by businesses
- Opacity: customers may not know what a credit buys, and rate changes can feel like hidden price increases
- What does a typical hybrid pricing model look like?
- A seat price with an included usage allowance plus overage charges
- A free product with no limits
- A one-time perpetual licence
- A price that changes randomly each month
- Why should vendors study the distribution of usage rather than just the average?
- Averages are too hard to calculate
- Regulators require usage histograms
- A small group of heavy users often drives most of the cost
- Distribution determines the product's name
- Which engineering choice can reduce cost per action?
- Always using the largest available model
- Removing all usage limits
- Doubling output length
- Routing simple requests to smaller models
- What lesson does the guide draw from reactions to AI coding tool pricing changes in 2025?
- Customers never notice pricing changes
- Transparency about what is included and advance notice matter as much as the pricing structure
- Usage pricing should be abandoned entirely
- Prices should only ever go down
Keep testing yourself
More quizzes picked for your level.