Across large enterprises, AI adoption is shifting from experimentation to operational pressure, with customer service often emerging as the first system that must work at scale.

The difference between pilots and production is no longer theoretical — it shows up in resolution rates, satisfaction scores and revenue impact. Zendesk Inc., a customer service software company, sits directly inside this shift, where automation, workflow redesign and data quality now determine whether AI improves outcomes or simply adds complexity. That tension naturally leads into how leaders are thinking about value creation and accountability, according to Tom Eggemeier (pictured), chief executive officer of Zendesk.

“I know the MIT study that was largely quoted in 2025 said 95% of C-suite are not getting the value from AI. We’re seeing the opposite from our customers … they’re getting a lot of value,” Eggemeier said. “People and our competitors are giving that value as well, but they want to go adopt it even further. They want to go get even more accuracy, more quality, more personalization and more adoption with their customers. We’re seeing 2026 is really the year of adoption and their year to take … accuracy up, quality up and to take automated resolutions solving customers’ problems.”

Eggemeier spoke with theCUBE’s John Furrier during an exclusive CUBE Conversation at theCUBE’s Palo Alto studios. They discussed how AI-driven automation is transforming customer service from a cost center into a measurable engine for growth, satisfaction and operational scale. (* Disclosure below.)

How customer service is being redefined by AI execution

Customer service is no longer evaluated solely on efficiency metrics, but on how well automation unlocks higher-value interactions. Enterprises are discovering that eliminating low-complexity work creates room to reinvest in loyalty, personalization and revenue-driving experiences. This reframes service as a strategic function rather than a defensive one, Eggemeier explained.

“Some customers are definitely doing cost-cutting from customer service and AI,” he said. “A lot are saying, ‘I can go solve a lot of my easy-to-medium complexity problems and take that as-a-service dividend to go drive revenue, to go drive loyalty with my customers through personalization.”

Execution, however, is where many organizations stumble. AI in customer service requires workflow reinvention, change management and continuous data feedback, not one-click deployment. Teams must redesign how work flows between humans and systems while maintaining accuracy at enterprise scale, Eggemeier explained.

“It’s a change management journey, where people are used to doing workflows,” he added. “They’re used to doing tasks a certain way, interacting with their customers. You need to go kind of turn that on their head and reinvent workflows, do things in a different way. It’s no longer press one button and everything’s beautiful. You really need to go to work to get the automation you want.”

From resolution data to measurable gains

Data quality and learning loops now define whether customer service AI improves over time. Automation must learn from both successful resolutions and human interventions, feeding accuracy, empathy and personalization back into the system. This creates a compounding effect when done correctly, Eggemeier noted.

“Sometimes you’re going to have AI agents solve a problem and you want to put that into the system on how they’re solving it,” he said. “Sometimes a human agent is going to solve the problem and you want the AI to learn from that. We think really having an end-to-end platform to use the data, to keep getting accuracy better, to keep getting empathy better, to keep getting personalization better is really a differentiation.”

At scale, results become measurable and hard to ignore. Organizations pushing past 50% automated resolutions are seeing meaningful gains in customer satisfaction, speed and consistency, according to Eggemeier. That performance pressure is accelerating board-level expectations around ROI.

“I’m seeing a lot of our customers pass the more than 50% threshold of getting automated resolutions,” he said. “What I’m seeing is customers are able to go solve their customers’ problems with their business or consumers, and they’re getting higher customer satisfaction. I think companies are able to really do that effectively right now.”

Here’s theCUBE’s complete video interview with Tom Eggemeier:

(* Disclosure: Zendesk sponsored this segment of theCUBE. Neither Zendesk nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE