Customer service operations are at a pivotal moment. Anyone who has tried calling customer service a few times knows that the traditional model is frustrating and broken. Long queue times, multiple unsuccessful service attempts, and unfulfilled promises to call back are just a few of the jarring irritations that customers may endure. For their part, companies have tried for decades to balance the twin pressures of improving the customer experience and reducing operational costs. But the rise of AI gives companies an opportunity to rewrite the paradigm by offering unconstrained service that is proactive, instantly available, and highly personalized to each customer's situation. Many businesses rightly see the potential of virtual assistants, co-pilot tools, and other AI capabilities to deliver both efficiency gains and an enhanced customer and human agent experience.
The business case is compelling. Pioneers are working toward unlocking large value pools and aiming to achieve a long-term productivity uplift of 60% or more, short-term P&L effects of 10% to 20%, and a customer lifetime value increase of up to 30%. Nevertheless, many customer service transformations remain confined to simpler customer service use cases. The reason: most customer needs require actual problem solving, not just conversation. According to our latest survey of 180 customer service leaders worldwide, only 28% of companies have unlocked measurable business value from generative AI (GenAI) in customer service. To succeed in this area, companies need to take a holistic, value-focused approach. This is where agentic AI comes in.
From Support Tool to Autonomous Actor
Equipped with memory, reasoning, and autonomy, agentic AI systems can perform real-time decision making and orchestrate end-to-end journeys, evolving their approach over time. Their distinguishing characteristics include three key abilities:
- Observe. Gather data from past interactions and real-time user inputs via application programming interfaces (APIs) to build holistic context from access to memory and system-wide signals.
- Plan. Analyze data to evaluate possible actions on the basis of defined roles, prior knowledge, and context; exercise reasoning capabilities via large language models; and set goals.
- Act. Execute planned tasks directly across tools and platforms, coordinate with other agents, and engage humans only when necessary.
This triad transforms AI agents from passive assistants into active operators that can learn, decide, and deliver at scale. However, the resulting system’s effectiveness depends critically on how the organization makes data and information available to AI agents. Beyond structured data and APIs, organizations must expose tacit knowledge—the insights and heuristics embedded in human expertise, frontline experiences, and customer interactions—in ways that AI agents can interpret and act on. Organizations can provide this support by having the AI agents shadow human experts and extract tacit knowledge from direct customer discussions, playing this into the overall knowledge pool and AI-supported unstructured data mining. Building this feedback-rich knowledge layer is a prerequisite for achieving sustainable autonomy, enabling AI agents not only to execute but also to continuously learn, refine their reasoning, and adapt to new contexts.
Rethinking the Service Value Chain
Customer service optimization consists of four stages: pre-empt, self-heal, self-help, and support response. (See the exhibit.)
Most AI deployments begin at the support response end, reducing average handling time, guiding agents with next-best-action suggestions, or summarizing calls. Although agentic AI increases value across all four stages, its highest value typically takes the form of pre-empting issues and enabling companies to self-heal problems.
AI agents can communicate with each other, prevent issues before they arise, automatically resolve back-end errors, and orchestrate and fulfill processes across the entire service value chain. The benefits include fewer customer contacts, improved customer satisfaction, and greater long-term productivity.
For example, agentic AI can detect billing errors before a customer notices them, autonomously investigate discrepancies, coordinate fixes across finance and support systems, and confirm resolution—all without having to raise a customer ticket. And this isn’t futuristic: the technology already exists.
Case Examples: Impact in Action
Organizations at the forefront of GenAI and agentic AI implementation are beginning to report significant gains. For example, a global technology company has achieved a 50% reduction in time to resolution for service requests, a European financial institution has automated 90% of its consumer loans, a Chinese insurance company has improved its contact center productivity by more than 50%, and a European energy provider has improved customer satisfaction by 18%. These success stories demonstrate that with the right strategy, GenAI and agentic AI can deliver sustainable improvements in customer experience, loyalty, and operational excellence, in addition to lowering costs.
Why Most Companies Still Struggle
Despite the potential of GenAI and agentic AI, many companies still struggle to unlock full value from them. The obstacles aren’t just technological but also organizational and strategic.
One hurdle is the scale of change management required. Not surprisingly, 98% of executives acknowledge that robust change management is crucial to the success of GenAI and agentic AI initiatives, but 50% identify it as the primary barrier to obtaining full value. This aligns with our project experience, where success with GenAI and agentic AI depends not just on deploying new tools, but also on redesigning processes, retraining teams, rethinking KPIs, and reshaping incentives—especially as we move toward agents that interact independently.
Another obstacle is the current fragmented technology ecosystem. Roughly nine in ten customer service leaders report difficulties in navigating vendors and system integrators. Moreover, GenAI and agentic AI solutions require robust orchestration across platforms, systems, and vendors, as well as customization of capabilities and use cases that are difficult to achieve.
Despite the challenges, companies have an opportunity to recover from setbacks. Among surveyed customer service leaders, 95% believe that GenAI will transform their business, and 78% expect it to scale within the next 24 months. To course-correct, successful organizations acknowledge existing gaps and reset their AI transformation strategy and ambition, progressing toward AI-centered process and organizational redesign and focusing on people enablement.
The Strategic Blueprint for GenAI and Agentic AI
Unlocking value from GenAI and agentic AI requires a disciplined, business-led transformation. Our experience from work with our clients and insights provided by more than 180 experts and customer service leaders worldwide yield five critical lessons:
- Make it business-led. Treat AI adoption as a strategic transformation, not just a tech deployment. True impact starts when business goals—not algorithms—drive the agenda.
- Focus on value. Prioritize initiatives based on business cases and value potential, not technological novelty. Every use case should include measurable outcomes.
- Get the tech stack right. Strong data and intelligence layers are essential for orchestration across systems and vendors. Without this foundation, even the best models will fall short.
- Build smart, scalable solutions. Winning organizations strike the right balance between buying and building modular tools, tailored integrations, and reusable components.
- Reinvent processes from scratch. Avoid the temptation to automate flawed human workflows. Instead, redesign processes for a world in which AI actively drives outcomes.
Crucially, companies must evolve their operating models to reflect a fundamental shift from AI as a tool to AI as a collaborator. As AI agents take on more responsibility across workflows, organizations must redesign how work gets done, rethinking team structures, decision rights, and oversight. Managing AI agents alongside separate human teams will become standard practice, requiring the leaders of customer service organizations to balance a mix of AI and human agents and create agile, outcome-focused teams. In this new setup, employees may become supervisors of AI, guiding and scaling its impact. To support this model, new roles such as conversational AI and agent designers, AI performance specialists, and AI process engineers will emerge, ensuring that human-AI collaboration is efficient and value-driven.
Winning with Agentic AI
Agentic AI isn’t just a tool. It’s a strategic capability that enables customer service to evolve from a reactive cost center to a proactive value creator. Realizing this potential requires more than deploying new technology. It demands a shift in how companies operate—upskilling human agents and leaders, transforming frontline workflows, orchestrating data and systems end-to-end, and embedding AI across the entire customer service value chain.
Achieving this transformation means redesigning processes, retraining teams, redefining KPIs, and introducing new roles to build, shape, and govern AI. Leaders must tackle structural change by building clean data foundations, embedding dynamic orchestration layers, and rethinking demand management to prevent the need for customer-initiated contacts before they happen.
The future of customer service lies not in the solitary actions of people or machines but in smart partnerships between the two. Companies that invest now in aligning people, processes, and technology around GenAI and agentic AI will not just improve service—they will redefine what customer service can contribute to the business, unlocking its full potential as a driver of growth, loyalty, and long-term impact.
