For the past two years, the AI race has been about one question: who could build the agent that knew every answer?
It made sense. Businesses needed proof AI could answer customer questions, automate repetitive work, and resolve increasingly complex requests. The first generation of AI delivered exactly that, transforming customer service in the process. AI assistants are now helping teams summarize conversations, draft responses, automate repetitive work, and resolve more customer issues than ever before.
But as AI becomes embedded into everyday service operations, the conversation is beginning to change. The challenge is no longer building one AI agent that knows every answer. It's building specialized AI agents deployed where they create the most value. The next race in AI won't be about broader capability. It'll be about deeper specialization.
“AI agents can act but too often they don't coordinate. They do one step. But the next step still starts at zero."
Great service has always relied on specialists
Service leaders already understand why specialization matters. The best service teams never expected one person to do everything. A billing expert handled payment disputes. A technical specialist solved complex product issues. A customer success manager stepped in when relationships mattered most.
Until now, AI has largely taken a different approach. The ambition has been to build one assistant capable of supporting every interaction. That made sense while the industry was proving what AI could do, but as Shashi Upadhyay, President of Product, Engineering and AI at Zendesk, explains, "AI agents can act but too often they don't coordinate. They do one step. But the next step still starts at zero."
That's the limitation of general-purpose AI. Customer service isn't a collection of isolated tasks. It's a connected system of people, workflows, and customer journeys. Solving one step well isn't enough if the next interaction loses context and starts from scratch.
And customers don't care who's solving the problem. They care that the right expertise appears when they need it and that the experience feels effortless from beginning to end.
The rise of the AI specialist
Rather than relying on one assistant to support every interaction, businesses will increasingly deploy specialist AI agents designed for different parts of the business. Some will specialize in technical support, edge case handling, and operational tasks like serial extraction. Others will focus on employee service, finance, healthcare, or retail. Each will develop deeper expertise because it is grounded in the knowledge, workflows, policies, and context unique to the work it is designed to do.
This is more than a shift in what AI can do. It is a shift in what a workforce can be. As Zendesk CEO Tom Eggemeier puts it: “We believe every business will soon run on specialized AI agents that work alongside human experts as one unified team. These agents will be more than just code; they will be team members, held to the same high standards of accountability as any human.”
The implication for service leaders is clear. The challenge will not be choosing between human and AI support. It will be thoughtfully designing how different kinds of expertise work together across the customer journey.
The next generation of AI will coordinate across other agents, business systems, and human experts to move work forward without losing context along the way. The customer may never see that coordination, but they will feel the difference in every interaction.
Service teams will blend human and AI expertise while delivering a unified, rich experience still feeling coherent, accountable, and personable to their customer.
Commerce is showing us what specialization looks like
Commerce is one of the first places where this shift is becoming visible. Customers don't think in pre-purchase and post-purchase service interactions with a site. They think about shopping. A single conversation can begin with discovering the right product, continue through checkout, move into delivery updates, and finish weeks later with a return or exchange.
Many AI systems divide that journey across separate agents. One helps customers shop. Another takes over after purchase. The result is fragmented context, unnecessary handoffs, and customers having to repeat themselves whenever the interaction changes.
That is why service organizations need more than one general-purpose agent to do it all. Different service tasks require different knowledge, business rules, tools, permissions, and levels of autonomy. A focused agent can be grounded in the right information, connected to the systems it needs, and evaluated against a clearly defined job—making its behavior easier to test and govern. Industry Agents can handle common, industry-specific work, while Custom Agents take on the policies, decisions, and workflows unique to each business. Working together across agents, systems, and people, these specialists can move the customer journey forward with greater focus, control, and expertise.
Specialization also helps compound learning as expertise improves over time. Through Zendesk’s Resolution Learning Loop, service interactions generate insights that help teams identify gaps, strengthen relevant knowledge and workflows, and improve how agents handle future work. In this way, organizations can continuously refine each agent’s role as they learn what works—and where more expertise is needed.
Commerce won't be the last place this happens. It's simply one of the first industries showing us what specialist AI looks like in practice.
Managing AI starts to look a lot like managing people
As specialist AI agents become part of everyday service delivery, the role of service leaders begins to change. The challenge is no longer deciding whether AI belongs in customer service. It's deciding how expertise should be organized across a team.
Which requests are best handled by a commerce specialist? When should a technical specialist step in? Where does human judgment create the greatest value? How should work move between specialists without disrupting the customer experience? These are questions service leaders have always answered when managing people. Increasingly, they'll answer them for AI too.
That's why the future of service won't be defined by one brilliant AI agent. It will be defined by teams of specialists—human and AI—working together toward the same customer outcome. Zendesk calls this the Autonomous Service Workforce, a coordinated network of specialist AI agents and human experts designed to resolve work end to end.
These teams will manage AI agents the same way they've always managed people: by matching the right expertise to the right work, orchestrating specialists behind the scenes, and creating customer experiences that feel effortless from beginning to end.
