Agentic AI represents the next evolution of artificial intelligence, following traditional network-based AI and Generative AI (GenAI). Distinguished by its decentralized, cooperative agents capable of independent operation and communication, agentic AI offers unprecedented potential to enhance operational efficiency and customer experiences. For Internet Service Providers (ISPs) and Managed Service Providers (MSPs), this technology could significantly impact bottom lines, provided they navigate associated risks such as vendor lock-in, cybersecurity threats, and privacy concerns.
What is Agentic AI?
Agentic AI refers to systems built around autonomous agents interacting through common protocols, resembling a digital ecosystem. Each agent independently performs specific tasks yet collaborates seamlessly with others, enhancing efficiency across complex processes. This modularity allows developers to easily manage and upgrade individual components, streamlining operational and customer-facing applications.
AT&T is at the forefront of this movement, implementing agentic AI to automate tasks such as fraud detection, network optimization, and software development. AT&T currently operates over 410 generative AI agents in production, developed through a framework known as Ask AT&T Workflows, which is integrated into its software development lifecycle.. The company has developed autonomous assistants that analyze customer accounts in real-time, providing staff with immediate options to address customer needs. These assistants utilize various large and small language models, selected based on task requirements, to perform their functions.
While mobile operators have been among the most proactive in exploring agentic AI's potential, ISPs can equally leverage its capabilities. They stand to gain significantly from adopting these technologies, particularly by automating routine management tasks, enhancing customer interactions, and differentiating their services through personalized, dynamic network offerings.
The role of the home gateway
For ISPs, agentic AI promises substantial operational efficiencies. It can potentially lead to Level 5 network automation, drastically minimizing the need for human intervention in network management. Some operators, for instance, aims for zero-touch network operations, leveraging agentic AI to optimize network performance, predict issues, and automate responses, significantly reducing operational costs and enhancing reliability.
An AI-driven home gateway can intelligently manage network resources in real time—dynamically adjusting QoS to prioritize a family’s video call or gaming session, or even responding to a simple user request to optimize the network for an upcoming task.
This also enables new context-aware capabilities; imagine simply texting your router to prepare for an important online meeting and having it automatically fine-tune the Wi-Fi network in response. Moreover, a gateway with embedded AI could learn from user behavior to deliver personalized content (tailoring services and media streams to individual preferences) and perform predictive troubleshooting by analyzing usage patterns and device logs to resolve issues before they disrupt the use. In essence, the once-humble router transforms into an intelligent home service hub.
Embedding agentic AI in home gateways can transform these devices from passive broadband pipes to intelligent network managers, managing the home automation systems such as thermostat and door locks, automatically prioritizing bandwidth for high-quality experiences during critical tasks like gaming or video conferencing. Such targeted Quality of Service (QoS) not only enhances customer satisfaction but also provides ISPs and MSPs with competitive differentiation, potentially leading to increased market share and customer loyalty against the giant OTTs like Amazon and Google.
Adoption Challenges
However, the transition to agentic AI is not without challenges. A significant concern highlighted by universal operators such as Vodafone who offer, fixed, mobile and TV services, is vendor lock-in. As agentic AI platforms like Google Cloud’s A2A (Agent-to-Agent) protocol are proprietary, ISPs risk becoming heavily dependent on single vendors, potentially limiting flexibility and increasing long-term operational costs.
Privacy and cybersecurity also remain significant concerns. Agentic AI’s decentralized nature, involving numerous interconnected agents exchanging sensitive information, can increase vulnerability to cyber-attacks and data breaches. ISPs and MSPs must invest robustly in cybersecurity infrastructure and ensure rigorous adherence to privacy standards to protect both customer data and their operational integrity.
Conclusion
In conclusion, agentic AI represents both an opportunity and a risk for ISPs and MSPs. To fully realize its financial potential, service providers must carefully balance the benefits of innovative AI-driven solutions with prudent management of vendor relationships and stringent cybersecurity practices. Those who navigate this balance successfully are likely to emerge as market leaders, benefiting significantly from the transformative capabilities of agentic AI.