AI-103 Practice Question 1: Batch vs Standard Deployment
Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 1 · question 1 of 10
A support organization runs one job every night that sends roughly two million closed tickets through a chat model to classify them by root cause. Nobody is waiting on the output interactively; the classifications only need to be in the reporting database by the start of business the next morning, and the team wants the lowest possible token cost without starving the quota that their live customer-facing assistant uses during the day.
Which Microsoft Foundry deployment type should the nightly job use?
Show answer and explanation
Answer:
Global Batch is built exactly for this shape of work: you send a large number of requests in one file instead of one at a time, and the service processes them asynchronously at 50% less cost than Global Standard with a 24-hour target turnaround. Critically for this scenario, Global Batch requests draw on a separate enqueued token quota, so the nightly job cannot disrupt the online assistant's capacity. The only thing you give up is real-time responsiveness, which this job does not need.
- A is wrong because Global Standard works but pays the full pay-per-token rate, and the nightly burst consumes the same quota pool the daytime assistant depends on. - B is wrong because provisioned throughput reserves capacity continuously. Sizing PTUs for a job that runs a few hours a night means paying for reserved capacity that sits idle the rest of the day, which is the opposite of the stated cost goal. - D is wrong because Data Zone Standard is a data-residency choice, not a cost choice. It is still pay-per-token with no batch discount, and no residency requirement was stated.
Reference: Understanding deployment types in Microsoft Foundry Models