AI cost attribution tools

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-tools-AI-cost-attribution-tools-creators

1. A company notices their monthly AI bill has tripled but can't tell which product feature is causing the increase. What problem does cost attribution solve in this scenario?

   1. It identifies which team or feature is driving the increased costs
   2. It charges the AI vendor less money
   3. It automatically reduces the spend by optimizing calls
   4. It replaces the billing system entirely

2. Which of the following is a capability of AI cost attribution tools?

   1. Deciding which teams should receive chargebacks
   2. Tagging each API call with team, feature, and customer identifiers
   3. Eliminating the need for any billing infrastructure
   4. Automatically reducing AI spending without trade-offs

3. A product manager wants to know how much the search feature costs compared to the chat feature. What must be true for this comparison to be possible?

   1. The costs must be below a certain threshold
   2. The finance team must approve the comparison
   3. The AI model must be the same for both features
   4. Each feature's API calls must be tagged with the feature identifier

4. Which statement best describes what AI cost attribution tools cannot do?

   1. They cannot tag API calls with metadata
   2. They cannot decide chargeback policies without human input
   3. They cannot roll up spend into dashboards
   4. They cannot distinguish between different customers

5. A finance team wants to create a chargeback model where each department pays for its own AI usage. What role does the AI attribution tool play in this process?

   1. It decides which departments should be charged
   2. It generates the invoices and collects payment
   3. It provides the data showing each team's spend
   4. It prevents departments from using AI

6. You are designing a call tagging schema for an AI-powered writing assistant used by both a free tier and a paid tier. Which dimension should definitely be included in the tags?

   1. The time of day
   2. The user's age
   3. The weather at the user's location
   4. The customer's subscription tier (free/paid)

7. A dashboard showing AI spend rolled up by team shows the 'Unknown' bucket at 30% of total spend. What does this indicate and what should be done?

   1. Many API calls are missing required tags, and tagging should be enforced at the SDK layer
   2. The company should stop using AI
   3. The teams are intentionally hiding their spend
   4. The AI tool is malfunctioning and should be replaced

8. What is the primary purpose of rolling up spend in dashboards for cost attribution?

   1. To hide individual transaction details
   2. To store data more efficiently
   3. To provide high-level views that enable decision-making across teams
   4. To generate invoices automatically

9. When designing a tagging schema, which of the following would be considered a valid attribution dimension?

   1. The development environment (dev/staging/prod)
   2. The color scheme of the UI
   3. The random seed used for generation
   4. The keyboard layout of the device

10. A product team notices their AI feature is driving significant costs but wants to keep the feature because it drives user engagement. What does this demonstrate about cost attribution?

    1. Cost attribution is only useful for cutting costs
    2. Cost attribution enables informed decisions about trade-offs
    3. Cost attribution automatically removes expensive features
    4. Cost attribution increases costs further

11. Why is it important to enforce tagging at the SDK layer rather than relying on manual processes?

    1. Manual processes are more accurate
    2. It ensures all calls are tagged consistently without relying on individual developers
    3. The SDK layer is faster than other methods
    4. The SDK layer is free

12. A startup is building a multi-tenant SaaS product where different customers have different AI usage patterns. Which attribution approach best supports billing each customer accurately?

    1. Using a single account for all customers
    2. Billing all customers a flat fee
    3. Tagging each API call with the customer identifier and rolling up by customer
    4. Only tracking monthly totals

13. The marketing team and engineering team both use the same AI feature but want to understand their separate costs. What must be implemented to make this possible?

    1. Two separate AI models
    2. Different pricing for each team
    3. A monthly budget for each team
    4. Team-level tagging on API calls

14. A company implements cost attribution and discovers the 'Unknown' bucket is growing. What is the most effective fix recommended in the lesson?

    1. Implementing stricter policies without technical changes
    2. Ignoring the unknown bucket
    3. Enforcing tags at the SDK layer
    4. Removing AI features

15. Before implementing cost attribution, a company cannot have conversations about AI spending because they only see one total bill. What does attribution make possible?

    1. Automatic bill payment
    2. Eliminating the need for finance teams
    3. Conversations about which teams, features, or customers drive spend
    4. Reducing AI capabilities