Throughout my career, I have been fortunate to have worked with major telecom companies across two continents across five different countries. This experience has led me to observe a set of challenges to telecommunications, unlike other industries, including:
- Persistent cycles of the latest generation (3G, 4G, 5G, and SOON 6G) upgrade networks.
- Large CAPEX is in demand with delayed or unclear ROI.
- Excessive disruption in core revenue (e.g. WhatsApp, Imessage, Tiktok, Zoom, YouTube, Netflix).
- Arpu is rising and decaying.
- Changing customer expectations, such as customers demanding a seamless, digital-first experience, including fast networks such as 5G, Fibreoptic Broadband, IoT integration, and value-added services.
In the face of these evolving challenges, the global communications sector is experiencing radical changes. Traditional carriers have once been set to operate as service providers focused on voice and data transmission, but are now rethinking themselves as a comprehensive technology provider.
This transformation represents more than just a rebranding exercise. Carriers are expanding their capabilities beyond basic connectivity services, challenging cloud computing, IoT solutions, artificial intelligence and digital platform services. They invest heavily in innovative customer experiences that reflect what software-defined networks, edge computing infrastructures, and what pure play technology companies have to offer.
This evolutionary driving force comes from changing market dynamics, increasing competition with digital native companies, increasing customer demand and increased demand for seamless integrated technology solutions, as well as fast demand and expectations. As the line between telecommunications and technology continues to blur, telephone companies need to position themselves not only as network providers, but as key partners in their customers' digital transformation journeys.
The transition from telco to techno has a major impact on CAPEX and OPEX due to different business models and investment priorities. While Technos has traditionally invested heavily in physical infrastructure (CAPEX), Technos focuses on software, services and digital platforms (OPEX). Currently, telecoms need to reassess and adopt an Opex-centric model to compete with spending patterns. This transformation extends beyond infrastructure and market mobility strategies to cover the entire customer service experience. The UK telecommunications division aims to provide a comprehensive overhaul of how businesses interact and support their customers.
Thanks to the important evolution of generative artificial intelligence, the way agent AI functions is clear. The agent AI located in the next major leap position does not just automate basic queries or generates general responses. It promises something much more ambitious for carriers. It is the ability for AI agents to make complex decisions, handle the entire customer journey, initiate actions across the system, and function independently of direct human surveillance. Imagine AI actively solving broadband outage issues or managing SIM swap end-to-end seamlessly.
But how close are UK telephone companies and their channel partners to see this level of intelligent automation being implemented for interaction with customers? And even more importantly, what should customer service leaders do to prepare now? To separate the burgeoning hype from real reality, let's explore what Agent AI really comes with, where it fits today's AI landscape, and how UK communications organizations can strategically prepare them for the ultimate arrival in customer service operations.
The role of Agent AI
Before we dig into the timeline and preparation of Agent AI in UK communications, it is important to understand what Agent AI is and how it differs from the traditional automation that many telephone companies already employ.
Agent AI has the ability to act autonomously towards defined goals. This means getting initiatives, inferring through complex customer issues and making decisions with minimal human input. Unlike basic rule-based chatbots and narrow AI assistants that frequently annoy today's customers, Agent AI can adapt and manage sophisticated, dynamic tasks.
But what really sets Agent AI apart for the communications sector is inference, making complex decisions across different systems, acting autonomously throughout the customer journey and technical workflow, and examining their own results along the way. Consider not only answering billing queries, but also actively investigate conflicts, cross-reference multiple systems (CRM, billing, network status) and applying credits.
How close is Full Agent AI?
Headlines and industry vendor pitches aren't close enough to make you believe.
According to recent industry data, less than 1% of enterprise applications worldwide used Agent AI in 2024, but that statistics can be misleading. It's more helpful for UK telecom leaders to think about AI maturity at five different stages of an agency, or how AI systems can work.
1. Do as I say: Task-based automation, such as automating robotic processes, is already widely used throughout communication in applications such as automating data entry, handling common service requests (such as address changes), and managing back-office functions.
2. Please tell me about research: AI aggregates data and brings insights to the surface. This is already happening in the multimodal model on Telecom today.
3. I recommend to me: AI synthesizes the data and recommends that humans are still in the loop.
At this point, levels 1-3 are achievable, already providing tangible value and are relatively well established.
Levels 4 and 5 are where the true transformational promises of Agent AI exist, but they are still not enterprise-ready due to their widespread and complex deployments.
4. Please decide about it and let me know: Minimal Human Surveillance: AI performs decisions and provides reports. An example would be for an AI agent to autonomously apply known outage credits and report actions.
5. Do that and don't mind: A fully autonomous agent that runs end-to-end workflows. This is the ultimate goal. The AI agents who manage the full-service transition are proactively solving technical issues from detection to customer notifications without human intervention.
Issues in Agent AI Recruitment
There are several important obstacles that prevent Agent AI from making a full flight with UK telecom customer service.
1. Legal and regulatory concerns: When decision-making moves from human agents to machinery, accountability becomes a pressing and complicated issue. In the UK's highly regulated communications sector, who is responsible for autonomous AI agents making false decisions that lead to excessive billing, disruption of services, or data breaches? Demonstrating an auditable decision path for regulatory compliance is an important hurdle. Furthermore, ensuring fairness, transparency and explanability of AI-driven decisions is of paramount importance under the UK AI Ethics Guidelines.
2. Data Quality, Integration, Legacy Systems: Agent AI needs to access more than standard customer records. To act effectively, it requires seamless interfaces with complex and often different internal systems. Many UK telecom providers, especially those with decades of mergers and acquisitions, are working on the spaghetti junction of legacy IT infrastructure. Mapping, cleaning and manufacturing this vast and diverse data in an accessible and practical way for autonomous AI is a monumental effort that many have yet to address.
3. Real-time adaptability and dynamic environment: AI models are usually frozen at the time of training. Unlike human agents who constantly coordinate decision-making based on the very latest information, the current agent system lacks its important, up-to-date awareness. This freeze is a major limitation for completely autonomous operations as communication environments are extremely dynamic and constantly changing networks, updating services and evolving customer needs.
4. Customer trust and human empathy: Customers are becoming increasingly comfortable with digital channels, but complex or emotional customer issues often require human empathy and subtle understanding. We trust fully autonomous AI for important services like Telecom (especially when things go wrong). By unfolding, do that and don't bother me right away.
Preparing Agent AI
We are still more than a year away from the widespread deployment of Agent AI, but the smart move is to strategically experiment with what is already available. This means focusing on the early stages of Agent AI (levels 1-3). These can drive already measurable ROIs and increase the efficiency of customer service.
More importantly, future-oriented telecoms in the UK should begin laying the important foundations for Levels 4 and 5.
- Mapping customer journeys and service workflows: This is basic. Understand all the decision points that human agents are currently making, the myriad systems involved, and all the decision points that require precisely the data needed for each step. For carriers, this often means analyzing highly complex technical troubleshooting paths and changes to multi-step services.
- Implement robust business process management: A structured approach to analyzing and optimizing core customer service and operational processes represents the best opportunities for automation. This is not just about applying AI. It's about getting rid of existing, chaotic processes first, and then getting them ready for AI.
- Developing comprehensive AI governance principles: Establish clear ethical guidelines, accountability frameworks and audit mechanisms for AI adoption. This should be done with UK-specific regulations (OFCOM, ICO) and consumer expectations in mind, ensuring a responsible, transparent and auditable AI deployment.
Agent AI may not be fully prepared to carry out today's communications customer service operations, but strategic decisions made in 2025 will clearly determine how ready your organization is. There's a risk of preparing now or catching up later.
