Requesting Emergency Changes on the AI Platform

In urgent scenarios, such as system outages or critical vulnerabilities, emergency changes to the AI platform may be necessary. This guide outlines when and how to request such changes.

When to Request an Emergency Change

Submit an emergency change request for high-priority incidents when:

A critical AI system or application outage affects multiple users.
There's a security vulnerability posing significant risk.
A temporary fix is needed to restore service pending a permanent solution.

Emergency changes should be exceptional. Whenever possible, follow the standard change management process.

How to Request an Emergency Change

To request an emergency change, follow these steps:

Open a high-priority ticket in the AI Platform Service Desk with details of the issue, urgency, and proposed change (rollout, downtime, test, and backout plans).
The AI Change Manager will review the request urgently.
If approved, follow the instructions to schedule and implement the change.
The Emergency Change Advisory Board will evaluate next steps.

Final approval from the AI Director and business stakeholders is required, based on criticality and risk assessment.

Rules for Implementing Emergency Changes

During an approved emergency change, adhere to the following rules:

Assemble a qualified team to execute the change swiftly.
Prepare vetted rollback and backout plans.
Provide detailed communication regarding impacts and downtime.
Follow standard procedures as much as possible, including peer review and testing.
Document all actions during and after the change.

Emergency changes are subject to audit, so maintain thorough records. Request emergency changes when necessary to mitigate risks effectively.