Has anyone found a real escalation path beyond standard AWS billing support? AWS's support AI told me, "I've confirmed that your account is eligible for a refund," and instructed me to retry after June 4 because of the monthly billing cycle. I followed those instructions and opened a case on June 9.
The dispute concerns $527.04 in Fast Snapshot Restore charges. AWS says I explicitly enabled FSR from my authenticated console session, but I didn't know what FSR was or where its control was located. I deleted the gaming instance and three experimental volumes on May 4, yet FSR remained enabled on the snapshot and continued billing until May 31. Roughly $474 accumulated after the experiment had been deleted.
My available CloudTrail records do not show a volume being created from the FSR-enabled snapshot, and AWS has not identified one either. I understand that FSR is charged while enabled, even if no restore occurs. My objection is that AWS's AI gave me a definitive eligibility statement and specific instructions that made the issue sound like a billing-cycle delay.
After more than five weeks, support declared the decision final and threatened suspension, collections, and possible credit consequences. They claim a supervisor reviewed the case but will not explain who reviewed it, what evidence was considered, whether FSR was ever used, or why the AI gave me incorrect refund guidance. What escalation route has actually worked for reaching an independent billing reviewer?
4 Answers
FSR is attached to the snapshot, not the instance or EBS volume. Deleting the instance and volumes does not remove it, so the snapshot can keep generating FSR charges until the feature is disabled separately. A restore is not required for the charge to apply. AWS generally expects customers to monitor these settings and costs themselves, so that part of their position is technically understandable, even though the interface and cleanup flow are easy to miss.
The stronger argument is not that FSR should never have been billed; it’s that AWS’s own support system made a definite account-specific representation and gave you instructions based on it. Save the complete chat transcript, case timeline, billing records, CloudTrail evidence, and every response refusing to answer your questions. Ask in writing for a formal billing escalation and a specific explanation of the AI statement, rather than continuing to debate the technical behavior of FSR.
Publicly posting the case number on AWS’s official support discussion site may get more attention than ordinary billing replies, especially if you don’t have an account manager. Keep the post factual and redact account details. You can also ask for an account review through Trust and Safety, but there’s no guarantee either route will overturn the decision.
The AI response should be treated cautiously because generative systems can confidently give incorrect answers. That said, if AWS makes the AI the primary support entry point, it’s reasonable to expect the company to investigate and address a misleading, account-specific promise instead of simply declaring the output irrelevant.
That’s my position. I accept that I was responsible for not noticing the FSR setting sooner, but the AI didn’t say I might qualify. It said my account was eligible and told me to retry after the billing cycle, which I did.