AI is already transforming how customer support teams work, and has been doing so with chatbots since before large language models (LLMs) were widely available.
But what’s coming next is bigger than just faster answers or smarter chatbots.
Speaking to CX Today, John Burghart, Managing Director at Deepdesk, explained: “We’re entering the era of agentic AI, where AI isn’t just assisting agents, but beginning to act more independently, with humans still in the loop.”
Deepdesk sees this as an evolution happening in three phases:
- The Current State: AI Tools Assisting Humans
Most contact centers today are using AI to support their agents, not replace them. Tools like real-time knowledge assist, summarization, and sentiment detection are helping human agents work faster and more efficiently.
For example, if a customer asks whether their bike insurance applies abroad, AI can instantly search through policy documents and bring the answer to the agent.
“It’s fast, accurate, and removes the need for agents to dig through multiple systems during a call.”
In many cases, AI is even completing small tasks in the background, like updating addresses in Salesforce or retrieving package tracking info, while the agent stays focused on the conversation. The agent is still in control of what the customer sees and hears.
- The Next Phase: AI and Humans Working Side by Side
Some contact centers have grown more confident in the quality of AI outputs, therefore “we’ll continue seeing a shift with more and more collaborative models,” said Burghart.
In this current stage, AI handles more complex tasks, while agents focus on nuance, empathy, and higher-stakes interactions.
Think of it as a flexible partnership. Depending on the industry or specific task, companies are now deciding how much work the AI handles vs. how much stays with the human.
A financial services company may lean more cautiously, while a retail brand might move faster into automation.
