AI is changing the contact center, but not in the straightforward way early predictions suggested. The initial narrative focused on the potential to reduce agent headcount and cut staffing costs, as automation would resolve routine queries, leaving fewer reasons for customers to speak with a person. That outcome may apply to some tasks, but it misses a larger shift in the work that remains.
Tim Richter, Senior Director of Product Marketing at Five9, told CX Today that the headcount-reduction narrative overlooks a changing customer service environment.
“It misses that the nature of inquiries is changing. In terms of volume, emotional intensity, and complexity. The initial hypothesis that AI is going to reduce human headcount and agent populations would only be the case if volumes, sentiment, and complexity were unchanged over the past several years. Data reveals that these factors are amplified.”
As AI takes on more simple, repeatable interactions, the cases reaching advisors are increasingly complex, emotionally charged or high risk. The contact center is not disappearing; its role is becoming more specialized.
Customers are also better informed. They can use AI tools to understand policies, compare options, and challenge bills or service decisions before contacting an organization. This changes both the nature of the interaction and the level of expertise expected from the advisor.
For CX technology buyers, the question is whether the organization has the people, data, quality processes and routing capability to manage the harder interactions that remain.
Routine Automation Changes the Work That Reaches People
AI can be effective for well-defined, low-risk requests, such as checking an order status, updating an address, resetting a password or answering a common policy question. But the contacts that escalate to people often need more than information retrieval. They may involve a disputed financial decision, a sensitive healthcare issue, a vulnerable customer, a service failure or a customer who has already attempted self-service and is frustrated.
This is changing the profile of the advisor role, Richter said.
“As AI is resolving more contacts, naturally the types of issues that are reaching agents are more complex and require higher degrees of specialty.”
Many contact centers were designed around generalist roles, with advisors working across a broad range of common queries, supported by scripts, knowledge articles and supervisor escalation.
That model is becoming less sufficient in sectors where the consequences of a poor interaction are significant. Financial services providers, for example, may need advisors able to handle fraud, affordability, complaints or complex product questions. Healthcare organizations may need staff who can manage sensitive patient interactions while following specific processes and governance requirements.
“Agents are going from being generalists—which they were five years ago—to increasingly specialists because that’s what they’re called upon to do,” Richter said.
This has direct implications for workforce planning. CX leaders should not assume that reduced demand for basic services will translate linearly into reduced demand for people. They need to assess the complexity, risk, emotional intensity and the expertise required of the interactions being handed to human teams.
A contact center with fewer routine contacts may require deeper specialist capability, better coaching and more targeted routing.
The Next-Generation Advisor Needs Judgement, Not Just Speed
As contact volumes become more complex, efficiency measures such as average handling time can become less useful in isolation. An advisor may need more time to resolve a high-risk case properly. The customer may need reassurance as well as a decision. They may need help interpreting a policy, navigating a multi-step process or understanding what will happen next.
At the same time, “agents have access to more information than they ever had before,” Richter pointed out. Real-time guidance, customer relationship management (CRM) records, AI-generated summaries, interaction history and knowledge content can all support resolution, but they also create a new cognitive demand. Advisors need the ability to manage several information sources while maintaining focus on the customer.
“They have to be looking at not just the information coming from real-time guidance cards, but access to various knowledge bases and customer records from CRM. There are multiple different information streams that they have to navigate,” Richter said.
This means the skills agenda is changing. Contact centers need to develop advisors who can interpret AI recommendations, exercise judgement when confidence is low and recognize when the suggested next action is not appropriate for the customer’s circumstances.
They also need greater empathy. Customers who reach a human advisor after attempting self-service may already be frustrated, and the interaction can carry a greater emotional burden than it did when every customer started in a live queue.
“When human agents are on the line with a very frustrated consumer, they need a higher degree of empathy than was called for five years ago,” Richter said.
This creates a practical evaluation point for buyers, because agent-assist tools should do more than simply deliver more prompts, notifications, or content, instead helping help advisors identify the relevant information quickly and reduce the effort of navigating systems during a live interaction.

