Generative AI has emerged as the new big tech disruption and today companies in every domain are exploring ways to use the technology to their advantage. Software-as-a-Service (SaaS) giant Zendesk too is leveraging the power of generative AI to provide a better customer experience. 

Zendesk provides a range of customer service tools and supports ticketing systems that are commonly utilised by multiple brands across the globe to manage and streamline their customer interactions. These new LLM-powered solutions by Zendesk help call centre agents efficiently handle customer inquiries, track and prioritise tickets, and provide a seamless customer support experience.

“By integrating LLM features into our intelligence panel, we anticipate doubling agent productivity. The panel offers reply recommendations, and conversation summaries, and allows agents to finetune their responses, resulting in enhanced efficiency,” Cristina Fonseca, head of AI, Zendesk, told AIM.

A cautious approach towards generative AI

Based in San Francisco, California, Zendesk powers billions of conversations, connecting more than 100,000 brands with hundreds of millions of customers over multiple channels such as chat, email, telephony, messaging, social channels, communities, review sites and help centres.

So far, Zendesk has introduced generative AI for a range of use cases that address some of the most immediate needs like summarisation, expanding replies, and tone adjustment – all aimed at empowering agents to craft exceptional responses. “We are actively working to make sure we leverage these large language models in every single piece of the product,” Fonseca said.

However, cautious is how Zendesk wants to be in adopting generative AI, considering the potential risks involved, including security threats demonstrated in cases like Samsung, as well as the issue of hallucinations. “For Zendesk, establishing trust in AI and our systems is paramount. While we acknowledge that AI solutions may not always be entirely accurate, we are mindful of the associated risks. We are already working with OpenAI to close some of the gaps that still exist. And in the meantime, we are not going to send our customer data to OpenAI unless they explicitly say that it’s okay to do so,” Fonseca added.

Currently, Zendesk is using GPT-4 as well as GPT3.5; however, they are not limiting themselves to these LLMs alone. “We are looking at other LLMs, models and companies whose technology we can leverage to develop the best possible solutions for our customers.”

AI before ChatGPT

Zendesk has been leveraging the power of AI even before ChatGPT came into the picture. Last year, the California-based company integrated AI into its Intelligent Triage, which is a customer sentiment and intent functionality. It automatically identifies the intent, language, and sentiment of customer tickets, providing valuable insights for effective handling and response.

By automatically prioritising responses based on customer intent, it empowers customer service teams to efficiently address and resolve customer issues, leading to improved productivity and customer satisfaction. “Our intelligent triage is basically pre-trained machine learning models that understand your customers like no other solution. And then we are building a couple of applications on top of it,” Fonseca said.

Will AI replace agents?

AI has the potential to automate certain tasks and enhance the efficiency of call centre operations. The technology is set to only get better with time, and it opens up the possibility of AI completely replacing call centre agents. Fonseca believes it is possible technically.

“However, in reality, for AI to replace call centre agents, companies would need to have well-integrated internal workflows that allow bots to seamlessly interact with automation systems,” she said.

But as things stand, many systems are not properly connected, hindering the full potential of automation. Moreover, customer service often involves decision-making, which is an area where human judgement and expertise are still crucial. “While there may be certain cases where AI can handle a significant portion of tasks, in most scenarios, we are still a long way from complete replacement,” Fonseca said.

Generative AI and customer experience 

After the release of ChatGPT, most organisations are exploring ways to have a ChatGPT-like bot to offer a better customer experience. Zendesk already operates in a very cluttered space with players such as Freshwork, HubSpot, Salesforce, and Zoho Corporations among others, going full throttle with generative AI.

“In the next couple of years, no one will buy a piece of software that doesn’t have AI and intelligence embedded. In today’s highly competitive landscape, with numerous SaaS companies emerging everywhere, the winners of the future will be those that not only provide effective procedures and systems but also focus on empowering the individuals behind those systems to excel in their roles,” she concluded.