Zendesk adds AI tools in pursuit of autonomous service
Zendesk makes the case for an autonomous service workforce for customer and employee experience.
Zendesk previewed a handful of AI agents, copilots and admin tools today that will roll out later this year. More importantly, the company gave a sneak peek at the bigger roadmap picture: autonomous service workforce management tools to supplement human employees and customer experience workers.
Copilots include tools for analyzing content, service trends and the root causes of ticket generation. Agents to deliver employee service -- built out of technology from Zendesk's Unleash acquisition last December -- operate in apps such as Slack and Microsoft Teams.
Also in early preview are Agent Builder, a no-code designer; Action Flows for AI agents, a workflow builder; Context Graph, the foundation for agent analytics reporting; and MCP Client, a model context protocol connector for Zendesk agents to tap outside data sources. MCP Server will come later, as will Quality Score, which measures the service quality of both human and AI agents.
Zendesk hopes the mix of tools will help get customer service teams from where they are now with their current IT stacks and support technologies to an autonomous service enterprise, said Colin Murphy, chief customer officer at Zendesk. But because users' boards and C-suites are also pushing for accelerated AI adoption to keep up with their competitors, they have to show quick wins with prebuilt, out-of-the-box agents that automate simple tasks, too.
An early adopter's service AI roadmap
Technology vendors want to make AI adoption as frictionless as possible -- in the shortest time possible -- but many customers still grasp for that "autonomous AI" goal, realizing it takes not only technology but also people and process management that are simpatico.
The business of booking monthly rentals -- a market that sits in a sweet spot between the short- and long-term stay -- has boomed since 2020, said Zendesk customer Cami Nariño, COO at Furnished Finder. The company launched in 2014, specializing in rentals for traveling nurses, who are typically looking to spend housing stipends.
Since the start of the pandemic, Furnished Finder has expanded from 20,000 to 300,000 properties. It has found a much wider audience among consumers -- such as digital nomads -- and workers across numerous industries and the military. Another growth area is insurance housing for individuals and families displaced from their homes.
In the middle of its rapid expansion, Nariño said, the company moved off of a homegrown reservation system to Zendesk for its case management and help center.
As it scales up, Furnished Finder is looking to automate more of its service interactions, with an emphasis on helping property owners prep, list, describe and price their rental spaces, among other things. That takes a lot of content -- with AI overlaid to surface the right answers and suggestions at the right times.
Human customer service will always be the bedrock of Furnished Finder, she said, because the company's reputation is built on high-touch service, especially for the property owners and travelers using it for the first time.
"So far, we have enabled Q&A through Zendesk," Nariño said. "As users come to Fern -- our AI assistant -- the chatbot is going to help answer all of the transactional questions they have. The next phase that we're exploring is allowing for account lookups, and then account action -- taking that a step further and actually being able to say, 'Hey, it's [a particular customer] that I'm chatting with, and they need support with the calendar, so I am going to update their calendar for this specific listing.'"
Both tactical and strategic AI
That brings up the distinction between tactical, task-based AI tools and broader, long-term strategic AI initiatives. Customers like Furnished Finder want it both ways from their technology vendors.
"There is a world in which you can try to automate everything," Nariño said. "But if you move too fast, you could lose some of the key components of actually providing service -- the human touch and the relationship building -- that I think are extremely important for our business."
This AI implementation story is being repeated at many Zendesk shops -- and across the board -- in customer service, Murphy said. Contact centers need some fast, tactical improvements that quickly expand and automate customer service availability to 24/7 -- and hand over the simplest customer queries to AI agents, while saving the thornier issues for human agents.
But as they build their hybrid human and AI workforces, they will also need broader, strategic AI to enable humans to manage and route work, as well as AI to spot content gaps in knowledge bases and report on trends driving customer tickets that can be solved with automation, when the agents are given the right content.
That represents complex change management algebra. AI is evolving so fast that, in some cases, contact center leaders need a hand just figuring out where to start.
"Some customers are maybe a little bit less ready for AI than others," Murphy said. "They need to rethink their customer service workflows, and they expect us to help them. We take that redefined process, we build it into our AI agent flows, and we build it into procedures that are generated and followed in our copilot technology."
Don Fluckinger is a seasoned B2B technology journalist with over 30 years of experience, specializing in enterprise IT, digital experience and content management. As a senior news writer at Informa TechTarget, he delivers award-winning analysis that helps IT and business leaders navigate complex technologies to enhance customer and employee experiences. Got a tip? Email him.
