Modern AI technologies — such as generative AI (genAI), large language models, and agentic AI — have been heralded as the future of customer service. And deservedly so: A conversation with a genAI bot can be built out in a small fraction of the time, and genAI bots now speak in a more comfortable and human manner and have the access and intelligence to blow away experiences we’ve had in the past.

But despite the hype, many organizations are discovering that AI alone isn’t delivering the transformative results they expected. Why? The real issues are systemic, persistent, and undersolved for. Specifically, it’s the outdated systems, fragmented processes, poor knowledge, and organizational inertia that prevent AI from reaching its full potential. Our research, Customer Service Must Evolve To Unlock AI’s Full Potential, analyzes these challenges — and suggests what enterprises should do about them.

We Are In A Moment Of Stepwise Change

Adoption of advanced AI capabilities is happening in the contact center. But what we’re getting today are simple and safe tools that improve how we do the same things we have always done — not actually changing service:

  • Call summarization. Agents no longer need to write their post-call notes from scratch. Instead, AI can compose the notes and the agent can edit and submit them — saving a significant amount of time.
  • Improved analytics and quality management. Knowing sentiment, and the content of every call, is opening up exciting new doors for insight into contact center performance.
  • Agent assist tools. From smart suggested answers to next-best-action tools and more, there are ways that AI now helps agents during interactions.

These are all examples of capabilities that make contact centers more efficient, but they aren’t providing the promised tectonic shift that brands are hoping for. That change will come when self-service capabilities — powered by modern AI — become a reality.

The Rubber Doesn’t Meet The Road — The AI Promise Versus The Reality

Today, we unfortunately know more about AI deployments that fall short than outperform. Air Canada’s chatbot made headlines in 2023 for giving a customer incorrect refund information — and the airline was held liable. British Airways’ AI chatbot mistakenly canceled bookings and issued incorrect travel advice, leading to customer frustration and reputational damage. Plus, we don’t currently speak much about customer service use cases for AI beyond chatbots.

But the key takeaway here is that these aren’t just chatbot failures — they highlight a critical and often ignored fact: AI is only as effective as the systems, data, knowledge, and processes that support it.

  1. Fragmented Tech Stacks Limit AI Effectiveness

Today: The contact center tech stack is an eclectic mix of old, new, emerging, and tailored tools — all of which are being upgraded to be new and improved. And it’s becoming hard for buyers to know what’s new and how improved it actually is.

Many customer service operations still run on siloed platforms that don’t communicate well with each other. Furthermore, many contact center platforms and tools also have overlapping features and functionalities, making it hard to reconcile for the best experience outcomes. AI tools often struggle to integrate with this complex ecosystem, leading to inconsistent data, broken workflows, and poor customer experiences.

  1. Platforms And Features Sell, But Experiences Suffer

Today: Many organizations believe that they have access to best-in-class technology and features, but many of them also admit that they aren’t using these to their full potential.

Contact center platforms build features at an incredible pace but find it hard to do two things in particular: 1) Enable the adoption of these features and 2) enable access to meaningful data and insights that can improve customer experiences. Consequently, many organizations feel like they’re overinvested in platforms and features but fail to establish either the need or the ROI from features they’ve bought.

  1. Workforce Readiness And Change Management

Today: AI adoption isn’t just a tech upgrade — it’s a cultural shift. Many organizations underestimate the need to reskill their workforce and manage change effectively.

AI adoption in customer service isn’t just about deploying new tools — it’s a cultural transformation that demands new ways of thinking and working. Many organizations underestimate the complexity of reskilling frontline agents, who often lack clarity on how to collaborate with AI tools, which in turn leads to underutilized technology and frustrated teams. To truly benefit from AI, companies must foster a mindset of continuous learning, collaboration, and adaptability across their service operations. Plus, they must understand how AI will transform the customer service workforce.

AI Is A Tool, Not A Magic Wand

AI is a powerful tool that amplifies what’s already working — and exposes what isn’t. To unlock its true potential, customer service leaders should focus on modernizing infrastructure, capturing human expertise, reskilling the workforce, cleaning up data, and continuously benchmarking AI readiness within their organizations.

Our research, Customer Service Must Evolve To Unlock AI’s Full Potential, discusses these challenges and outlines best practices to ensure an AI-enabled future for customer service.

Visit my Forrester bio page (as well as those of my coauthors, Max Ball and Kate Leggett) and click “Follow” to receive notifications about upcoming research. Forrester clients can also schedule an inquiry or guidance session to delve deeper into this topic.