The rise of agentic AI in network infrastructure requires careful planning to reap the rewards, writes Gary Sidhu, SVP Product Engineering, GTT
As digital ecosystems grow in complexity, agentic AI offers a new way to optimise, secure and scale network infrastructure.
With the proliferation of connected devices, speed and responsiveness has never been more important. Agentic AI, which instead of merely responding to requests can operate autonomously and proactively within its environment, can help accelerate opportunities for turning insights into action.
Simply put, agentic AI listens, learns, and develops strategies capable of revolutionising how we work, especially in network operations where it shifts from reactive to proactive, improving resilience and security. It can automate network management, real-time threat detection, and traffic optimisation, enhancing efficiency, strengthening security, and boosting network performance for seamless and secure operations.
But how can it be implemented, where can it have the biggest benefit, what is the role of human oversight and what lessons can we learn from the introduction of agentic AI? In this article, I’ll cover these key points and give advice to businesses looking to harness its potential.
Agentic AI and infrastructure design
The successful implementation of agent-based AI systems requires careful planning.
Firstly, it is important to clearly define goals and key performance indicators for their use. Then, a major challenge is the seamless integration of the solutions into the existing IT network infrastructure. Training and operation of the systems also require the availability of sufficient and high-quality data. Finally, there are ethical considerations of implementing agentic AI that companies need to address from the outset, such as data privacy, protection, governance, human oversight and transparency, to ensure trust is built.
Agentic AI requires guidelines over which data it can access, from where, and whether it is able to share certain data externally. This is imperative to consider within an AI strategy to ensure both customer and organisational protection from data and regulatory breaches, particularly in the light of regulations such as the EU AI Act.
If your implementation plan takes these considerations into account, nothing stands in the way of the effective use of agentic AI. With digital agents, businesses can streamline their operations, meeting rising customer service expectations. A report by Gartner predicts that by 2029, AI will resolve 80% of common customer service issues without human intervention. These agents analyse customer sentiment in real time and provide tailored responses enhancing customer engagement.
Managing distributed networks at speed
Agentic AI is now playing a pivotal role in network infrastructure and cybersecurity, helping organisations move beyond traditional, rule-based systems. Unlike conventional tools that passively monitor and alert, digital agents can actively observe network behaviour, identify anomalies in real time, and take autonomous action to resolve emerging threats. This enables a faster response to incidents, reducing downtime, and therefore helps avoid costly disruptions.
Agentic AI is already being embedded across networking and security infrastructure to deliver real-time, measurable value. The NSaaS model (Networking and Security as a Service) is evolving into something more dynamic, where agentic capabilities enable smart routing, adaptive policy enforcement, and predictive resource allocation. These enhancements ensure better performance, greater visibility, and stronger protection for global customers operating in complex conditions.
Bridging the gap
There is growing demand for integrated cybersecurity and networking solutions from Cloud providers, with many organisations viewing this convergence as essential to enterprise resilience. In this setting, agentic AI offers a unique advantage; it blends machine learning with autonomous decision-making, allowing digital agents to adapt in real time while maintaining stable and efficient network operations.
This shift from static systems to intelligent, self-improving agents is reshaping how businesses think about their digital foundations. With this strategic mindset, early adoption of agentic AI gives network providers a chance to get ahead of the curve with smarter services, improved reliability, and a more personalised customer experience.
Rethinking operational efficiency
While we are still at the beginning of the AI journey and its potential is yet to be fully realised, McKinsey found that 77% of companies are either using AI or exploring its potential. It has already changed workflows, it still requires a level of human management, but agentic AI enables new possibilities.
It can become more than a support tool. It can become an active participant in business operations, freeing up resources and creating greater efficiency. In networking specifically, the benefits are becoming clear. While machine learning has been used for tasks like digital twins and anomaly detection, agentic AI can manage these processes autonomously. This reduces the need for human intervention at every step and enables networks to become more resilient, secure, and adaptive to real-time demands.
Nevertheless, learning and development around AI in the workforce remains a business imperative. Counterintuitively, while flawed data is often tolerated in human decision-making, we don’t have the same leniency with AI. According to Gartner, 30% of generative AI projects are abandoned after the proof-of-concept stage, primarily due to issues related to data quality, risk management or high costs, highlighting the difficulty organisations face in AI initiatives and importance of getting implementation right from the off.
For some companies, agentic AI could mark a shift from promise to performance – where AI becomes not just an experiment, but a business-critical capability aligned to strategic goals.
An autonomous, adaptive future
The path to resilient, scalable networks lies in intelligent automation. Agentic AI brings real-time learning and decision-making into operational workflows, helping manage complexity and optimise uptime. For organisations ready to move from passive AI to proactive systems, the future lies in intelligent agents – and the resilient, secure networks that support them. This is not just a technological shift, but a reimagining of what network infrastructure can achieve when paired with AI designed to think, act, and adapt.
Gary Sidhu is SVP of Product Engineering at GTT, where he leads the innovation behind the company’s global network platform. He oversees EnvisionDX, EnvisionCORE, and EnvisionEDGE – the backbone of GTT’s enterprise services.
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by visiting our LinkedIn page.
