VMware Uses Machine Learning to Add New Powers to Workspace One
Last week, VMware announced that they are launching a host of "smart" capabilities for its VMware Workspace One product that aims to deliver customer insights and improve security. By using machine learning (ML) and other technologies, the identity management platform will reportedly monitor user behavior, software performance, and hardware information. These features come during a time when an increasing amount of business applications are implementing artificial intelligence (AI) capabilities.
Intelligent Management
Before taking a closer look at the new capabilities, it's important to understand exactly what VMware Workspace One is. For the uninitiated, the platform lets IT departments remotely manage all employee devices and programs no matter where they are. This includes everything from mobile devices and virtual desktops to apps such as customer relationship management (CRM) platforms. In turn, the solution provides users with self-service access to their favorite apps via the VMware Workspace app on their machine, where they simply enter one set of login credentials for their programs. When we reviewed the VMware Workspace One solution last year, we were impressed with its features, awarding it our 4-star "Excellent" rating.
Workspace One Intelligence was the first of their newly announced capabilities for the platform. By using ML, the software will now monitor and harvest data from the customer's network to help spot opportunities for improvement. The platform also offers visualization tools to make understanding these insights simpler.
For example, let's say your company deployed a new app for employees and usage of the app is far below what you had hoped. Workspace One Intelligence, which is now able to track app performance, would tell you that most of your employees are using a slower, outdated version of the app. By using the new Intelligence capability, you now have the insights to spot the problem and instruct your employees to update the app. Workspace One Intelligence's features are available now.
VMware also announced the launch of its Workspace One Trust Network. Essentially, VMware has opened up the platform's application programming interfaces (APIs) to a select network of partners, including Crowdstrike, McAfee, Symantec, and other major security players. VMware has also allowed these companies to feed data into the platform. Let's say a security threat is present in your network that affects unpatched Windows PCs. In theory, these partners, with their wealth of knowledge in the latest cybersecurity threats, would be able to feed an alert regarding the unpatched systems into VMware Workspace One. You can even set up a rule to automatically push those kinds of updates.
"By leveraging the knowledge base of these partners, we are in a position where we are much better suited toward quickly identifying threats and solving those problems," said David Grant, Vice President of Marketing at VMware. Customers will be able to take advantage of the Workspace One Trust Network later this spring.
Smarter Business Tools
VMware's infusion of "smart" capabilities for VMware Workspace One is similar to what we've seen in other parts of the business software landscape. This month alone, Microsoft announced a new set of tools with so-called "intelligent features," including new capabilities for its Microsoft Power BI platform. AI chatbots are taking over our customer service software and AI-laden products are becoming more ubiquitous in the workplace.
Pam Baker is a business analyst, PCMag contributor specializing in analytics, and author of the book Data Divination: Big Data Strategies. Baker has spent a great deal of her career covering the world of Big Data and its impact. While the tech is certainly exciting, she argues that these "intelligent" initiatives are not without their hurdles.
"ML is only as good as its models and not all models are created equal," said Baker. "Furthermore, it's very difficult to test ML models because it's almost impossible to get the same output from two identical runs. And it's very difficult to modify, correct, or manage a model that you can't test." According to Baker, this is known in the research field as the "ML reproducibility crisis."
Regarding the security promises VMware and other vendors have made when it comes to ML, Baker was equally cautious. "Vendors often tout predictive capabilities for security purposes but take that with a grain of salt," she said. "ML works reasonably well with behavior analytics in predicting insider threats, and it works reasonably well in the prediction and early detection of known security threats. But it can't predict a security threat or vulnerability that is unknown. The key is to understand its limits but leverage its strengths."
To put it in more nerd-friendly terms, Baker used an interesting metaphor. "It's not Batman. That would be a dedicated security analyst," she said. "ML is Robin, the helpful sidekick. But Batman still needs his utility belt."
This article originally appeared on PCMag.com.