How Predictive Analytics Helps Minimize Device Downtime and Keep Users Productive

Applying artificial intelligence and machine learning to a massive pool of device health data enables services like HP Proactive Insights to predict failures and address performance issues.

Minimizing downtime and staying productive

Predictive analytics is being used in myriad industries to revolutionize maintenance processes, and IT is no exception. During the pandemic, it’s providing a much needed helping hand to IT in dealing with end-user devices that are now in employee homes, far from the corporate help desk.

Traditionally, machines of various types – from those on plant floors to a UPS in a data center – were maintained strictly on a calendar basis. That meant parts were replaced because a calendar said it was time, not because there was anything wrong with them.

Predictive analytics uses artificial intelligence (AI) technology to make maintenance smarter. Now tools are available that can detect when a machine or a device, including a computer, is not functioning as it should, according to an established baseline.

Using such tools, you can now predict when a computer is going to have a problem and take steps to fix it before the issue actually manifests in a breakdown – and a loss of user productivity.

Learning from 20 million devices

HP, for example, pools anonymized data from more than 20 million devices into a data lake, then applies AI and machine learning technology to find correlations, says Karl Paetzel, Director of Software and Service Product Management at HP. With a data pool that large, the company can establish with a high level of certainty what a “normal” level of operation looks like and flag any devices that stray outside the norm.

Through the HP Proactive Insights service, the company monitors key hardware characteristics, including thermal issues and hardware and software errors. It can also diagnose performance issues, such as when one customer saw a spike in users suffering the dreaded “blue screen of death.”

“That can be challenging to diagnose in a small device population,” Paetzel says. “Because we have a large data pool to analyze against, we can see if a certain pattern is happening beyond one customer.”

Sure enough, it found two drivers that were causing the issue and advised the company to change them. Once it did, the spikes went down.

Predicting failures, automating responses

Using analytics, HP can also predict when a component such as a battery or hard drive is in danger of failing. Laptop batteries have been a particular issue during the pandemic because they are designed to be continually charged and discharged. But users working from home tend to have them docked and charging all the time.

“We can detect when a battery is under continuous charge and not being conditioned properly,” Paetzel says. In some instances, HP may detect users don’t have certain settings turned on, such as HP Health Manager and important security settings. It can also detect if a machine’s BIOS is out of date, which can lead to security issues.

When it does find an issue, HP Proactive Insights can automatically kick off routines to get the problem addressed.

“We can display whether the device is under warranty,” Paetzel says. “If so, we can trigger a response based on the entitlement of that service.”

For large organizations, HP Proactive Insights can tie in with their IT service management (ITSM) tool to automate a response. The same goes for customers that subscribe to other HP services, such as HP Active Care for next-day response to problems.  

“If we can predict a failure, IT doesn’t have to take as many calls from users,” Paetzel says. “It’s a nice way to reduce costs.”

Just as important, users remain productive ­– and IT doesn’t have to lift a finger. 

Click here to learn more about how HP Proactive Insights can help you optimize your fleet of endpoints.


Copyright © 2021 IDG Communications, Inc.