Microsoft wants to keep things running smoothly in the cloud.\u00a0<\/p>\n
The company is seeking to patent a system for \u201cautomated incident triage\u201d<\/a> in cloud computing environments using generative AI to automatically track, route and handle any errors.<\/p>\n \u201cConventional approaches leverage general machine learning models to aid in triage and diagnosis,\u201d the patent notes. \u201cHowever, the performance of these approaches is limited due to a lack of domain knowledge in general machine learning models from various triage teams.\u201d\u00a0<\/p>\n Cloud \u201cincidents\u201d come in many forms that can undermine the health, performance and security of a cloud system, including bugs, outages, service errors and vulnerabilities. When an incident is detected, a triage engine will sort it into the right category to troubleshoot it properly.\u00a0<\/p>\n Each potential category has its own dedicated generative AI model trained on past data to properly handle such incidents. The models digest information related to the incident and make recommendations on which team should tackle it.<\/p>\n Cybersecurity has long been Microsoft\u2019s bread and butter. The company\u2019s patent history includes tech to weed out anomalies in cloud environments<\/a>, prevent AI \u201cjailbreaks\u201d<\/a> and evaluate data health<\/a>. It\u2019s only natural that security extends to Azure, which has become one of the company\u2019s biggest moneymakers as AI drives up demand.\u00a0<\/p>\n In the most recent quarter<\/a>, the company\u2019s cloud unit made $42.4 billion, up 20% from the previous year\u2019s quarter. Azure holds 22% of global cloud market share, according to CRN<\/a>, second only to Amazon Web Services.\u00a0<\/p>\n