AWS has launched a guided SageMaker Studio workflow for generative AI inference recommendations, moving from an API-led process to a visual interface under Jobs and Inference optimization. The workflow allows teams to describe their expected workload, choose an optimization goal—minimize cost, minimize latency, or maximize throughput—compare ranked configurations, and deploy an endpoint from SageMaker Studio. The flow begins with four workload choices: Interact, Generate, Summarize, or Custom, which lets a team supply a representative JSON Lines dataset from Amazon S3 plus concurrency and average output length. AWS introduced optimized generative AI inference recommendations in April, and the July release adds the guided Studio experience around that machinery. SageMaker benchmarks on real GPU infrastructure with AIPerf, and preset results are useful starting points, while representative custom traffic gives stronger evidence. Recommendation jobs have no extra service fee, but benchmark compute and deployed endpoints cost money, and model or traffic changes can invalidate the result.