Concerns About AI‑Driven Responses in the Standard $100/mo Support Plan

I recently opened a support request under the Standard support plan and noticed that most interactions are now handled by an AI agent. I’ve run into several issues that seem to indicate the AI is not yet ready to replace human‑driven support for paid plans:

1. Outdated guidance: The AI repeatedly referenced old Azure portal terminology (e.g., “Recovery Services” instead of the newer “Resiliency” area) and pointed me to UI elements that no longer exist.
2. Inability to parse provided logs: I sent diagnostic logs in my reply, but the AI did not analyze them. Instead, it instructed me to manually search the logs for specific text—something even basic AI tooling can do today. It asked this question

   1. In the JSON output you shared from the Azure CLI command, does the "properties.enhancedBackupAlerts" field equal "true"?

   .
   But I had just sent it the json output in the body of the previous email, so it should know the answer to that question.
3. Delayed follow‑up: Initial responses were fast, but the last reply never arrived. It has now been four days without an update on a Severity B ticket. If AI is going to triage responses, there should be a guaranteed human follow‑up when the AI cannot produce a timely or relevant answer.
4. Lack of backend insight: The AI seems limited to surface‑level instructions and cannot perform deeper investigation or access internal diagnostic information. It mostly redirects the user to perform all discovery and troubleshooting manually.

If AI is going to be used for paid support interactions, I strongly recommend addressing these issues to ensure the experience meets the expectations of customers paying for Standard support.