Zendesk AI: Level up with Generative AI for a more intuitive and intelligent CX platform

zendesk is expanding the use of artificial intelligence (AI) in its customer experience (CX) platform with the release today of a number of new features that the company is branding as Zendesk AI.

Zendesk’s new AI capabilities include advanced bots for handling customer inquiries across a range of industries. Customer service agents will now also benefit from AI-powered assistance in responding to inquiries. Additionally, the system integrates smart rating features that use sentiment analysis and intent detection to help route requests correctly.

While many of Zendesk’s AI capabilities were developed by the company’s teams, the new services also benefited from a healthy dose of generative AI through a partnership with Open AI. The generative AI component will scale across Zendesk’s AI suite, helping generate responses and summarize content.

>>Follow VentureBeat’s continued coverage of generative AI<

“We believe in making AI easy to understand, use and accessible to everyone,” Zendesk head of AI Cristina Fonseca told VentureBeat.

What Generative AI adds to Zendesk

Zendesk was no stranger to AI prior to its partnership with OpenAI.

Zendesk took a big step into AI in 2021 with the acquisition of Portuguese startup Cleverly AI, of which Fonseca was previously CEO. In September 2022, Zendesk announced a major update integrating Cleverly AI technology to speed up customer service responses. Fonseca said today’s release of Zendesk AI is an evolution of what the company announced in September 2022, which was specifically aimed at the retail sector. He noted that the new version is available to a larger group of customers.

Zendesk has offered bot capabilities for a while, but Fonseca said OpenAI’s generative AI foundation has greatly expanded those capabilities.

“We believe that software should be intelligent out of the box, and for most bots to work, you need to spend a lot of time training them,” he said. “This is one of the main features we’re rolling out now, with pre-trained bots that already understand customers.”

Generative AI will also help power a revamped set of capabilities for increased productivity of customer service agents. Zendesk AI uses OpenAI technology to support summation and sentiment analysis for requests which, in turn, helps agents respond effectively. The ability to assist agents in creating responses is also part of the new update.

Why Generative AI alone is not enough for CX

Fonseca emphasized that while Zendesk AI is using OpenAI’s generative AI, it isn’t abandoning its intelligent AI roots.

Intelligent triage, for example, is a key element used by the product that was developed by Cleverly AI to understand customer intent, sentiment and language to appropriately address a request. For that system, Zendesk has their own proprietary models that have been trained on Zendesk data. Fonseca said these models understand customer service because they have been specifically trained on customer service data and provide a high degree of accuracy.

“The way we see (OpenAI’s generative AI) is basically a tool to help us accelerate things that were already on the roadmap, and it makes a lot of sense to add them to our product suite,” Fonseca said.

For example, he noticed that Zendesk was developing its own approach to suggesting new responses to customer service agents, as well as creating content for a knowledge base on different issues. Without OpenAI, he said the content draft that Zendesk was able to generate wasn’t quite polished. Now, with OpenAI integration, Zendesk data can be used to generate well-written answers and knowledge base articles.

“We’re looking to leverage OpenAI and large language models (LLM) to help us refine everything we do, beyond content,” he said.

With all the power AI offers Zendesk and its users, Fonseca cautioned that it’s important to realize that AI can’t and shouldn’t do everything when it comes to customer experience. He noted that CX is complex and includes workflows, and sometimes systems, that aren’t integrated with Zendesk.

“I think the first thing we should do for our customers is help them understand what should be automated and what shouldn’t be automated,” he said. “Because if something can’t be automated, there’s no point in trying to get a bot to talk to a customer if the bot can’t add any value.”