AI-200 cheat sheet

Updated September 27, 2026

Last-minute reference for AI-200. The data domain is the largest; the other three are roughly a quarter each.

Exam facts

Duration120 minutes
Passing score700 / 1000
PriceSet per country or region; shown at booking
ReplacesAZ-204
Language focusPython

Skill areas

AreaWeight
Develop AI solutions by using Azure data management services25–30%
Develop containerized solutions on Azure20–25%
Connect to and consume Azure services20–25%
Secure, monitor, and troubleshoot Azure solutions20–25%

Requirement to service

RequirementAnswer
Rebuild the image when the base image is patchedACR task with base image trigger
Run exactly the tested imageReference by digest
Secret in a container on App ServiceKey Vault reference + managed identity
Canary release, instant rollbackContainer Apps multiple revision mode
Scale a worker on queue length, to zeroKEDA Service Bus rule, min replicas 0
Existing Kubernetes manifestsAKS
Read your own writes, cheapestCosmos DB session consistency
React to new documents in Cosmos DBChange feed processor
Vector search with SQL filters and joinsPostgreSQL + pgvector
Semantic cache, lowest latencyAzure Managed Redis vector index
One consumer per message, retriesService Bus queue
Every subscriber gets a copy, with filtersService Bus topic + subscriptions
Notify many handlers of an eventEvent Grid
Ordered processing per entityService Bus sessions
No credential stored anywhereManaged identity + DefaultAzureCredential
Same key, different value per environmentApp Configuration labels
Follow one request across servicesOpenTelemetry context propagation

Vector indexes

ServiceOptions
Cosmos DBflat (exact, small dimensions), quantizedFlat, diskANN (large scale)
PostgreSQLIVFFlat (needs data first), HNSW (works on empty tables, higher recall)
RedisSearch index with a vector field

pgvector operators: <-> Euclidean, <=> cosine, <#> negative inner product. The operator must match the index’s operator class.

KQL in one line

requests | where timestamp > ago(1d) | summarize percentile(duration, 95) by bin(timestamp, 5m) | render timechart

Join requests to dependencies on operation_Id.

Traps

  • Redis is a cache, not a system of record. Durable data goes in Cosmos DB or PostgreSQL.
  • CPU scale rules cannot scale to zero. Use an event-driven KEDA rule.
  • Default indexing indexes everything. Exclude unused paths, including the embedding path from the range index.
  • Messages that “disappear” are in the dead-letter queue. Something has to read it.
  • Lost Event Grid events mean dead-lettering was not configured.
  • Reading a secret by version ignores rotation. Read by name.
  • AZ-204 material covers topics that are gone and misses the vector content.

Night-before checklist

  • Name the four areas and which is largest
  • One sentence each on when to pick Cosmos DB, PostgreSQL or Redis
  • Service Bus versus Event Grid, cold
  • Check ID and proctoring rules; see exam day

Take the 20-question practice test and check your weakest area.