Time to converge networks, security and AI to beat the CISO blues

Many of today's digitally-focussed businesses are looking to extend high-performance network connectivity to a hybrid workforce, sometimes on a global basis. And they are having to do this in the face of a surge in security threats. At the same time, Generative and Agentic AI are rewriting the rulebook on how and where work gets done, creating the need to safeguard and connect a whole new class of AI-driven apps and services. CIOs and CISOs are expected to manage this three way challenge – networks, security and AI. The ideal way for them to achieve this is through some kind of converged solution that is designed to handle all three.

The idea of converged solutions is, of course, nothing new. Over the past few years, we have seen a coming together of the once separate worlds of networking and security. Solutions such as SASE have tried to meet the need for a more integrated approach. But surging levels of cyber risk are driving hard pressed network managers and CISOs to seek new and better answers. They must move on from a mess of point solutions in favour of a platform that offers optimal network protection while giving them full visibility of possible anomalies. 

AI is busily complicating the challenge further, argues Rik Turner, Chief Analyst, Cybersecurity with independent consultancy Omdia: "AI is changing the game once more, providing a hugely increased attack surface and a raft of new headaches," he says. 

It may be time, he believes, to incorporate AI into the mix, converging it with the existing SASE model. It's certainly time to move on from the kind of basic SASE envisaged by Gartner in 2019 when it first launched the concept of the secure access service edge: "Originally it was a bundle of technologies spanning the two domains – networks and security," says Turner. "With the explosion of the new variants of AI since 2022, first Generative then Agentic, it was inevitable that these new technologies would be applied to SASE, just as they are invading every other aspect of tech. AI can help speed path selection on the networking side of things, while in security it can streamline the detection of threats and enable faster response."

It's time, he suggests, for organisations to benefit from a form of SASE that supports their employees' use of software-as-a-service (SaaS) apps, given that a lot of GenAI is delivered in SaaS mode. "In this way, SASE can also be used to provide security and governance around those services," he says. "In other words, which services are sanctioned and which blocked, and within the sanctioned ones, what kinds of prompts can be used and what data can be up- and downloaded by the employee."

Bringing it all together

The kind of converged solution that can meaningfully address the triple challenges of networking, security and AI cannot just be stitched together from legacy products, says Aditya K Sood, VP Security & AI Strategy at Aryaka, a company specializing in unified SASE. The need for an integrated and converged approach to networking, security and AI is, he says, behind Aryaka's newly launched Unified SASE as a Service 2.0. The new platform incorporates several features to accommodate booming remote work and rising AI adoption, ensuring that users can securely connect to any application, anywhere, with performance, simplicity, and agility [see chart below].
 

"The attack surface is expanding," he explains. "And simplistic solutions are not going to help us. The route to protection lies in a solid network. The days of looking at alerts and logs in siloes is over, and what we need is a convergence point offering one place where you can look at any anomalies happening in the network. A place where you easily see what's happening with availability as well as security and integrity. This will give you a true picture of what your enterprise looks like. CISOs won't get this from a series of individual tools. With networking and security in their own silo, you don't get visibility of both sides. If both are contextualized at the same time you can get a true 360 degree picture. You need a platform to serve that purpose."

Integrating the networking and security part is one thing, he says, but bringing in the issue of AI takes it to a new level: "Use of AI is increasing exponentially, which means we need to understand the behaviour of users and how they are interacting with it. At Aryaka, we believe that you need one console for all this – network, security and AI. With a unified control plane you can enforce policies. You can design policies so that they go across similar data assets. This means you don't have to deal with individual policies for different kinds of asset." 

Visibility is key to any such platform, argues Sood, given that the CISO is not dealing with one standard threat model: "You need visibility that looks at more than one aspect at a time," he says. "If the CISO doesn't have this then they are in difficulty. They need a very definite taxonomy that defines, say, the different components of the AI ecosystem. What is the threat model for each component? From there you can start to build your defences and strategy. 

The need for accountability

A converged platform is the only way to offer today's CISO a way to confront a major difficulty: accountability. In the event of a security breach, the CISO will be called upon to report not only to their board level colleagues but also to the regulatory authorities. The visibility offered by a converged platform is their best weapon.

"There was a time when there would be an incident and the CISO would react in a certain way," says Sood. "With the world as it is now, you have to assume that the breach has already happened. The CISO must be proactive rather than reactive. When talking to regulators or other executives on the board, numbers are what they understand. Your fellow C-suite people will be asking about ROI for security spend. You need to be able to talk about the number of high risk vulnerabilities you have tackled over a particular period of time. This needs to be transformed into an exploitation matrix, showing how you've reduced the attack surface. The numbers have to speak.  It's a way to demonstrate the return on the tools deployed. You need a way to show how AI has increased the attack surface, perhaps by 4x. You have to say 'We are entering into the AI world and this is the risk we face. AI adds to the need for visibility, to see for example which users are bringing in their own devices and using AI on them." 

A fully converged approach to networking and security is the only way to help enterprises gain the performance, protection and visibility needed to confidently adopt AI at scale, in both on-premise and hybrid environments. The right converged offering allows organizations to innovate without exposing intellectual property and assets, and permits its employees to operate where they need to with all the tools at their disposal to enable maximum productivity.