Versa Networks Adds Generative AI to SASE Platform – Source: securityboulevard.com

Versa Networks this week revealed it has extended its secure access service edge (SASE) platform to include generative artificial intelligence (AI) capabilities that identify potentially malicious behaviors in real-time.

The Versa AI platform can now generate controls based on policies defined by cybersecurity teams and identify sensitive data and surface recommendations to improve data loss protection based on patterns.

In addition, the Versa Networking platform can now provide recommendations via a natural language interface using generative AI models, dubbed VersaGPT, that have been trained using the company’s documentation and knowledge base. At the same time, there is now an AI model, dubbed VANI, that makes use of machine learning algorithms to identify anomalies and preemptively reroute traffic in real-time.

The company has also extended the predictive AI capabilities it provides using machine learning algorithms to enable file and code snippet preprocessing. That capability enables the identification of malware without having to create a separate sandbox and reduces the time and effort needed to investigate suspicious activity.

Cybersecurity teams can also now use machine learning algorithms to automatically isolate threats into microsegments in real-time to limit the potential blast radius of a breach.

Kumar Mehta, chief development officer for Versa Networks, said this capability uses large language models (LLMs) that the company has embedded within its VersaAI platform alongside the existing AI models the company previously developed using machine learning algorithms.

The generative and predictive capabilities are built using a data lake that Versa Networks uses to build multiple types of AI models, he added.

There’s no doubt that AI is transforming the way cybersecurity and networking will be managed going forward. The goal is to enable IT teams to both respond to and thwart threats in real-time in addition to being able to predict how a cyberattack might be anticipated and prevented, noted Mehta.

Time, as always, is the enemy of cybersecurity. The challenge is that as cybercriminals also embrace AI, attacks are only going to increase in volume and sophistication. The only way to combat those threats will be for cybersecurity teams to invoke AI capabilities to identify and respond to threats in real-time. Today, most cybersecurity teams require minutes, sometimes even hours, to investigate malware that wreaks havoc the longer it is allowed to run.

Naturally, there’s a lot of skepticism concerning the level of protection that can be achieved using AI. Given the chronic shortage of cybersecurity expertise, there is little doubt organizations will need to rely more on AI to level the playing field. The issue will be absorbing the cost of acquiring new platforms at a time when more organizations than ever are sensitive to IT costs.

Of course, the cost of a breach can easily be measured in millions of dollars; in comparison, the cost of acquiring a platform infused with AI capabilities might be trivial.