AI · Article
KT and Microsoft Cut Network Fault Detection From 10 Minutes to 2-3 Minutes With AI Agent
인공지능신문 ·
At a glance
KT and Microsoft built an AI agent platform called AIO-Net that finds network problems much faster. The AI learns from veteran engineers' experience to help analyze causes, and people make the final decisions.
Key points
- AIO-Net detects network anomalies in 2 to 3 minutes, down from over 10 minutes.
- The AI agent writes cause analysis reports in about 5 minutes on average.
- The AI can only check and analyze devices, and it has no permission to send commands to them.
- KT plans to expand the platform from 70 core network devices to 280.
Why it matters
This is a good example of using AI in critical infrastructure while people keep final control. It may also offer a reference for power grids and data centers..
Original source
KT·마이크로소프트, 네트워크 문제 감지하는 AI 에이전트 개발…장애 탐지 10분에서 2~3분으로 단축
인공지능신문 · Nov 12, 10:41 KST
Read on 인공지능신문 ↗AI summarized the original article. Check the original for accuracy. Copyright belongs to 인공지능신문.
Other coverage
Microsoft · Ranking and more news
More from this field
- AWS Quick Adds Chat-Built Flows, Updated Web App, and Teamwork Among Agents
AWS Quick added updates that let a team turn one person's work routine into a shared, repeatable process. Flows can now be built and changed entirely in chat, and they can run from the web, desktop, or mobile app. Agents working together can also give more complete answers.
Amazon AWS Machine Learning blog · Nov 21, 10:06 KST - Amazon Bedrock Web Search Adds Five Languages and Country-Based Results
Amazon's Web Search tool for Bedrock, which gives AI agents cited, up-to-date web results, now supports Spanish, French, German, Italian, and Japanese, plus non-US English. A new optional country field can also steer results toward sources from a specific country. Existing customers get the language update automatically, and pricing stays the same in every language.
AWS What's New · Nov 21, 09:17 KST - AWS Adds Shared Cache Skill to Cut Self-Hosted LLM Cost and Latency
AWS wrote about a shared cache that lets self-hosted AI models reuse saved work across servers, aiming to cut cost and wait time. The post shows how a coding agent can help set it up. Amazon says it can cut inference cost by up to 40% and first-token delay by up to 80%.
Amazon AWS Machine Learning blog · Nov 21, 09:14 KST - Can AI Really Find Black Friday TV Deals? A BBC Test Says Not Yet
A BBC team tested four popular AI chatbots to see if they could find a good Black Friday deal on a 55-inch TV. The results were mixed, with many links broken or out of date.
BBC News · Nov 21, 09:06 KST