How Zendesk builds the future of AI-powered service

How Zendesk builds the future of AI-powered service

The News

On April 15 to April 18, 2024, Zendesk held its annual Relate event, including a half day analyst track on April 15. The event was attended by around 1,600 customers, partners and analysts.

It was about Zendesk’s strategy, which revolves around – no surprise here – AI to deliver better customer experiences. As part of this strategy, Zendesk also made clear how the past twelve month’s acquisitions of Klaus, Ultimate, and tymeshift get integrated into Zendesk’s customer service offerings, enriching and rounding them off. The company is betting big on AI, working on the assumption that interaction volumes between customers and companies are continuing to increase very fast. As a conclusion of this, service needs to become AI driven to accommodate this scale. Secondly, Zendesk sees AI as the technology underlying the necessary high degree of personalization. Together, this is estimated to increase the market size available to CX solutions that automate CX labor tremendously.

At the event, Zendesk had three key announcements. They were

  • AI agents to improve self service solutions,
  • a copilot that helps agents solve incoming tickets faster and provides insight to further optimize the service and
  • a workforce engagement solution that helps improve the productivity of digital and human agents as well as the quality of conversations.

Behind all this lies the recognition that customer service is very much conversational.

Customers and partners that I talked with had a keen interest in learning more about AI use cases. Many of them had started to use AI but estimated themselves still in early stages. 

The bigger picture

The customer service software market has become very crowded with players ranging from service specialists to suite players, with classic call-center companies making inroads, too. Additionally, there is a raft of niche companies that are offering specialized add-ons to improve customer service solutions, from connecting them to workforce management or -engagement solutions, to QA solutions and many more. It is quite easy to see a kind of a consolidation going on here.

A good number of these solutions use advanced statistics or machine learning based technologies to increase the overall efficiency of customer service and/or to achieve a higher degree of automation. If not outright embedded into each other, solutions are regularly integrated using APIs, with customers often choosing to work with a small number of vendors in a kind of an ecosystem approach. With service centers being regularly trimmed to efficiency, companies usually follow one of two base strategies: they either look for efficiency, aka cutting cost, or for offering better service without adding cost. The latter is often combined with setting up the service center as a profit center, as opposed to a cost center.

The latter strategy was exemplified by a Zendesk Relate customer that built a continuous improvement strategy around its ticketing system. Three of the main KPIs that are measured and shall be maximized are customer satisfaction, employee satisfaction, and the ratio of automated responses. The company is achieving its goals using generative AI to solve issues, another AI to predict a customer satisfaction score. Human agents are involved if the AI cannot generate a response or if the predicted customer satisfaction for this response is not high enough. In this case, human agents are tasked with improving it. Quite the same happens with agent responses: The agents’ responses are evaluated for predicted customer satisfaction and sent back to the agents for improvement if the predicted customer satisfaction score is not high enough. This is combined with an automated coaching system that agents are encouraged to use to continuously improve their own productivity. The results of implementing this strategy – which admittedly is a bit more complex in terms of systems, metrics, and culture – are quite amazing.

My point of view and analysis

Let’s cover two topics here, the event itself and the content.

Starting with the event.

Looking at it through a customer lens, it was chock-full of information and, more importantly, the ability to talk to experts about the roadmap and to see working software in the exhibition hall. The customer presentations were very interesting.

Keynotes are always an interesting thing. Robert Richman very relevantly and excitingly talked about customer service requiring a service culture. Having Michelle Obama on the first day was certainly interesting. On the other hand, no one I talked to found much relation to Zendesk in her story. This is the eternal challenge that I have also seen at other events. While it is important to work with “luminaries”, it is sometimes difficult for the audience to relate their stories to the story that the vendor tells.

From an analyst perspective, I would have wished to have more access to company executives to dig deeper into information. Still, the ability to speak to customers and partners was valuable. 

Big kudos to the organizing team that mastered the challenge of creating an event that was interesting for customers, partners, and analysts.

Coming to the content.

As said in the beginning, these days every CX strategy involves AI, if it doesn’t outright revolve around it. This includes Zendesk. From a capability as well as a vision capability, Zendesk seems to be positioned better than many of its competitors in the customer service arena. And there are certainly a good number of competitors. Zendesk’s position is a result of a good product mix, the early creation of “ready-to-use” AI models to improve service efficiency as well as of interesting acquisitions. Of course, it is important to deliver this vision fast. These days, most advantages tend to be of temporary nature.

The good number of competitors is not only because the customer service market is crowded. It is also partly attributable to the fact that Zendesk claims to offer a “complete CX solution for the AI era”. This broadens the competitive field significantly, also into areas that Zendesk does not cover or does not cover with a strong product. CX is not limited to customer service.

On the other hand, Zendesk’s mission is to “power exceptional service for every person on the planet”. This focus on service processes somewhat reduces the ambiguity of the terms “CX” and “CX department”. As one of the execs said “we are great when it comes to tickets” – which is true. In short, the company maintains a focus on customer (and employee) service. It is important for customers to keep this in mind.

Having said that, Zendesk also offers a sales solution, Zendesk for sales. It was notably absent throughout the event; yet is important for the complete CX story. The same is true for the Sunshine platform, although in this case there still was a lot of talk about the platform, without naming it. This makes a lot of sense, both in terms of the need and also a need to make the fresh acquisitions part of this platform. 

Talking about these acquisitions, they are important add-ons to the core service functionality. This is especially true for the QA capabilities that come with Klaus and the AI engine that is delivered by Ultimate. Tymeshift as a WFM solution plays out its strengths increasingly with bigger growing service centers.

Service processes will be automated more and more to achieve and maintain high levels of customer and employee satisfaction. This needs a solution that has the necessary intelligence. Zendesk is on a good way to get this job done. It already now solves parts of this problem and has many of the remaining parts of this solution at hand – and is working on combining them to a whole.