Mist Unveils AI-Driven Wireless Support Assistant

Mist has debuted the artificial intelligence (AI)-driven Virtual Network Assistant (VNA) for wireless operations, with an integrated help desk module. It will be available via its channel partner network of resellers and managed service providers.

Powered by Mist’s AI engine, Marvis, VNA is a new cloud-based micro-service that uses natural language processing (NLP) to make it easy to query the Mist global cloud for real-time monitoring of mobile client activity. VNA uses data science to easily identify Wi-Fi issues, understand the impact of wireless problems, correlate events across the wireless/wired/mobile device/IoT domains and auto-alert on anomalies.

“VNA is the next step in Mist’s journey towards building an intelligent AI-driven network that simplifies operations, lowers operational expenses and gives unprecedented insight into the wireless user experience,” said Bob Friday, CTO and co-founder at Mist. “We started with a robust distributed micro-services-based software architecture built on a cloud-based platform that collects and manages an enormous amount of data. On top of this, we implemented a patented methodology for organizing and classifying this data into domain specific service levels. Mist now delivers a VNA that can answer questions on par with a wireless domain expert.”

VNA brings NLP to network operations so IT staff can easily understand their network and client environment without having to manually sift through a myriad of data in numerous locations. Types of queries include:

  • Why is Bob’s smartphone having a problem?
  • Were there any anomalies between7 a.m. and 9 a.m. on the main campus?
  • List the three sites with lowest performance.
  • How many clients are on the guest network?

Issues it detects include:

  • DHCP (duplicate addresses, server down, …)
  • RADIUS (wrong user name, expired certs, …)
  • WAN (packet loss, intermittent dropping, …)
  • WLAN (interference, coverage, roaming, …)
  • Security (Pre-shared key typed incorrectly)