Как Cognizant помогает телеком-операторам внедрять агентов ИИ в реальном бизнесе

Today, agentic AI is a powerful driver of telco transformation. By automating complex tasks and decisions, it can reduce operational workloads, improve customer experiences, and create capacity for innovation and growth. Yet a gap remains between promising use cases and enterprise-scale deployment. Scaling agentic AI often requires telcos to redesign workflows, modernise infrastructure, strengthen orchestration, and integrate agents into business decision-making. Success also depends on the telco’s unique operating environment. Moving from pilot to production requires the right context — built around purpose, process, and people. Finding the right purpose Agentic AI promises benefits across multiple areas, but the challenge lies in understanding where it can deliver the greatest value for an individual telco. While it may be tempting to pursue AI projects that deliver smarter interfaces or operational assistant bots, identifying pain points and untapped opportunities helps telcos build a robust business case before deployment. This not only optimises upfront investments in AI infrastructure, but also ensures a stepwise approach to rollout that is better aligned with existing teams. This lets telcos take a more prudent approach — focusing resources on the gaps that matter most. The result: a stronger return on investment, including churn reduction, revenue uplift, operational savings, and faster service provisioning. In a customer experience example, identifying which interactions are best suited for agentic conversation and transaction automation helps telcos evaluate the right business case for agentic AI. The structure of a deployment would differ between an offloading of simple order change requests — where an agent can independently handle requests and approve changes — and a high-stakes customer journey such as resolving a long-standing billing dispute, where human expertise remains essential, but an agent can provide guidance and support. Redesigning the process Most enterprises view agentic AI as a change in itself; however, it is more aptly characterised as an enabler of change. By redesigning the processes and workflows it is deployed to support, telcos can move beyond adding autonomous capability to transforming core operational models. For example, agentic AI can be deployed in marketing to assist with recommending new mobile plans. Without a process redesign, an AI agent may be bolted on to an existing customer chatbot, allowing it to recommend plans when prompted and execute requested changes. By redesigning the workflow, however, the agent could continuously analyse customer usage patterns, predict when subscribers are likely to benefit from a different plan, and proactively engage them. Instead of waiting for customers to initiate conversations, the telco creates an always-on engagement model that not only automates upselling but also transforms in-base monetisation. These redesigns are necessary not just to extract the most value from an agentic AI initiative, but also to ensure that gaps, broken workflows, and fragmented systems are addressed beforehand, creating a stronger foundation for scalable deployment. Putting people in the loop While agentic AI may navigate the field alone, humans set the rules and step in when needed. Good governance is essential for agentic AI deployments, but this requires more than establishing guardrails; humans must remain involved in the decision-making loop. Much like redesigning business processes, preparing an AI-ready workforce is critical to maximising the value of any agentic AI deployment. While the vision of agentic AI may be end-to-end workflow autonomy, the reality is that human teams must work alongside agents to achieve operational impact. This goes beyond humans existing at the administrative layer. Instead, telcos should embrace a collaborative model. Agents support human decision-making, and humans step in to complete or refine transactions an agent has started, for a more personal touch. For this, agentic platforms need to be designed for seamless handovers and coordination, with human teams and the agents themselves trained to work hand-in-hand. A good example of this is fraud management. An agentic AI system can autonomously detect suspicious activity and execute immediate actions such as temporarily suspending a service. Human analysts, on the other hand, can then review these cases in detail, exercise contextual judgement, and implement a more proportionate and informed resolution. How Cognizant bridges the ambition gap Ultimately, the success of any agentic AI initiative lies in the context surrounding its deployment. For technology leader and AI builder company Cognizant, leveraging this context to design tailored agentic AI solutions is key to bridging the gap between AI infrastructure investments and business value. For telcos, Cognizant’s Communications, Media, and Technology (CMT) practice focuses on accelerating digital transformation across OSS/BSS, AI-driven operations, and network automation. Its core focus, however, goes beyond integrating AI solutions; it is about building the right capabilities for telcos to transform their operations. “We are at our core an AI builder company,” says Vijay Anand, Senior Vice President of Cognizant. “We don't simply integrate AI into telcos; we build it. If you look at the CapEx investments that telcos have made, they've spent billions of dollars on AI infrastructure. Our job is to take the value that resides in that infrastructure and translate it into business outcomes, whether that's network operations, business operations, or other strategic growth areas.” The company’s approach is centred on ensuring AI deployments are not only embedded natively within enterprise environments, but also designed for long-term scalability. From personalised AI concierge agents that enhance customer engagement to autonomous BSS agents that streamline operations and service delivery, Cognizant’s suite of agentic AI solutions enables CSPs to deploy and orchestrate intelligent agents across the telecom value chain. Real-world outcomes While most agentic AI deployments remain in their early stages, Cognizant's CMT practice has already supported several telco AI transformation programmes across the globe. Just recently, the company worked with a leading broadband provider to relaunch its customer-facing chatbot using a hybrid model that balances deterministic and agentic AI. The business impact was immediate: first contact resolution improved by 17% within a year, while increased context awareness drove an 8x rise in chatbot usage. A continuous feedback loop has also enabled the provider to implement more than 500 improvements. Self-service now resolves more than 30% of certain customer requests. This frees live agents to focus on complex, personalized support, while the voice bot has achieved a 65.2% resolution rate of its own. “Agents are easy to build, but enterprise-grade telco agents are not,” adds Vijay. “It's not about buying the model; it's about owning the harness. That’s the true moat that an enterprise should retain. This includes not only reimagining the business process but also building the evaluations, security, observability, guardrails, trust, governance, and designing for human-in-the-loop from day one, and that’s where we come in.” Delivering real value By aligning agentic AI to the right purpose, redesigned processes, and empowered people, telcos can move from experimentation to enterprise-scale impact. For Cognizant, bridging this ambition gap is what transforms agentic AI from an impressive demonstration into a scalable enterprise capability, one that enables telcos to modernize operations, elevate customer experiences, and unlock new opportunities for growth.