Where are the blueprints coming from?
Pick a source, then the Cloud Architect agent will assess the loaded blueprints.
Pick one or more. Each gets its own shape recommendations and indicative TCO so you can compare side by side.
Auto picks a current-gen family per host by workload (compute→C4, general→N4, memory→M3, tiny/cost→E2). N2 is comparable to AWS m6i / Azure Dsv5 and is sustained-use-discount eligible (~30% off). E2 is cost-optimized but not SUD-eligible.
How to bundle VMs into migration waves. Application groups keep a business app's hosts together; Subnets keep each network segment (portgroup / inferred /24) together for a subnet-by-subnet cutover.
Manual lets you set how many VMs land per wave. Let the AI decide hands the sizing to Your Cloud Architect, which proposes a count you approve before it's applied.
How Your Cloud Architect should steer this engagement — it biases the 7R framing, wave sizing, and the pre-flight de-risk questions. Balanced is the neutral default.
Free-text directives the advisor follows. Advisory only — they never override required safety gates (rollback, blast-radius, uptime SLA).