Written by Stephen Amstutz, Head of Strategy and Innovation, Xalient
Today most organisations are thinking about or deploying AI and, in effect, trying it out. This is supported by Gartner, which states that approximately 80% of enterprises will have used generative artificial intelligence (GenAI) application programming interfaces (APIs) or models by 2026. As AI drives value for organisations, it is fuelling further demand and adoption. One way that organisations are using AI is to review large data sets to identify trends and patterns so they can sequence data accordingly. Today there is a phenomenal amount of data we can now use to train AI. OpenAI’s ChatGPT and other large language models have further enabled vendors to introduce AI into their products to enhance the user experience. AI can engage data that lives in the online world in a more natural way, capturing certain activities and events to provide more information and context.
The importance of accuracy and context with AI
The challenge is around how these large language models dig into complex subjects. There is a lot of superficial data and there will be limitations for LLMs. For example, they can only look at data that is available. In other words, LLMs predict the best next word based on what has already been provided. This is great if you want help with anything text-related. However, accuracy and context are important, because the results generated by these LLMs, and other AI solutions, will drive business and security decisions and therefore must be accurate.
Over the past year or so we have seen a plethora of new AI capabilities coming to market. However, AI is a bit of a double-edged sword. AI solutions are attracting equally close attention from threat actors who are realising that – while they can be used by companies to identify security weaknesses and address them – they could themselves be a weakness in a company’s security. So, while AI presents breakthroughs in the ability to process logic differently, it also blurs the lines between humans and machines. This is why identity is crucial to ensure that organisations can securely connect people to technology.
Many vendors are launching new AI solutions
Many of our Identity and Access Management (IAM) partners have announced new advanced AI capabilities that address this issue without introducing AI risk. For example, built on a foundation of AI and ML, SailPoint’s Atlas platform delivers the right level of access to the right identities and resources at the right time. This provides visibility, insight, and remediation, so organisations can adapt and ensure the security of every identity’s access.
Okta AI is a suite of AI-powered capabilities that empower organisations to harness the power of AI to build better experiences and protect against cyberattacks. And it is not just the identity vendors, Zscaler’s AI-powered SASE connects organisations more securely than SASE based on traditional SD-WAN. Traditional networks were designed to solve yesterday’s problems and they introduce unnecessary complexity and risk. Today’s hybrid workforce and cloud-first apps need a new connectivity capability built on zero trust principles and this is where our partner, Zscaler, can help.
While all these innovations are moving the needle in terms of how they are enabling organisations to harness the power of AI, secure identity, deliver faster secure networking and harness new connectivity capabilities, they are all point solutions working in isolation. For true visibility, they need a layer of observability and orchestration to rapidly effect change and enable customers to ‘close the loop’ and reduce the amount of friction involved in making changes.
The importance of AIOps
Today businesses can’t tolerate operational delays and service disruptions; organisations are turning to Artificial Intelligence for IT Operations (AIOps) solutions to help resource-stretched IT teams. These solutions enable automated processes that free valuable resources to pursue innovation while AIOps solutions help organisations maintain uptime, reduce manual incident-management tasks, spot anomalies, and increase productivity.
Here at Xalient our AIOps capability, MARTINA, is an observability and orchestration layer that augments our partners’ AI technology and effects change. Each piece of infrastructure MARTINA monitors is a vantage point that acts like a camera angle on traffic flow and helps us to determine both the usage and performance of the network and makes sense of the data across the multiple vendors we support. It contains an API ingestion engine delivering visibility across all these point solutions with context and improvements for the end user experience. Context is key because MARTINA doesn’t duplicate what other technology partners’ solutions are doing, it leverages or augments these by pulling the right context, at the right time, and correlating this across the board.
Delivering self-healing connectivity
In effect, by using a combination of telemetry, AIOps, and machine learning, organisations can create self-healing connectivity and an infrastructure that is responsive, adaptive, and ultimately delivers better availability. This ensures high levels of resilience and performance are maintained, enabling customers to deliver a consistent, positive user experience across the organisation’s network. MARTINA lets our customers know exactly what alerts they need to be prioritising and be more proactive in monitoring and resolving impact assessments.
Ultimately AI will change our world as we know it, and there are many positive productivity gains and innovations that will result. But there is a need to understand context and, in a world where IT skills are in short supply and in-demand, the ability to automate IT operations swiftly to ‘close the loop’ on all these point solutions is essential. In my next article, I will tackle how organisations can leverage AI to provide secure self-healing connectivity in a zero trust world.
