How Zendesk Intelligent Triage steps up the customer service game

0
93 views

Share on LinkedIn

The News

On September 14, 2022, Zendesk announced the release of its new customer sentiment and intent functionality: Intelligent Triage and Smart Assist. These new AI based solutions shall “enable businesses to triage customer support requests automatically and access valuable data at scale.

Intelligent Triage and Smart Assist are the next step in Zendesk’s vision to create accessible CX AI for companies of all sizes. The technology uses proprietary industry expertise and insights from trillions of customer data points and applies a vertical lens. This creates models custom to each business capable of identifying the intent, language and sentiment of each customer interaction.

This unique approach to applying machine learning creates more personalized and informed interactions to better serve customers. For example, specific inquiries, such as “I’m having problems with payment”, can be automatically sent to an agent who is equipped to handle billing for a quicker resolution, while inquiries that include language written in all capital letters or in a sarcastic way will indicate a highly negative sentiment and be routed to the top of the queue.

The new capabilities include:

  • Instantly route and prioritize revenue drivers, ensuring agents are working on business-critical requests
  • Analyze distribution of requests so businesses can better plan operations, collaborate across departments and identify improvement opportunities supported by data for more efficient CX operations
  • Automatically guide agents on how to best resolve a customer’s issue in real-time, understand context, recommend solutions, and improve coaching and training with valuable insights
  • Continuously boost accuracy as the AI solutions receive feedback on predictions and recommendations
  • Detect sensitive information automatically to meet compliance and security needs or extract confidential data like names, addresses, phone numbers, usernames, and financial info for use in workflows

The bigger picture

Customer support is a very difficult business. All too often agents are not trained too well and service departments suffer from high turnover. On the other hand, customers expect to solve issues faster and faster. This drives a demand for efficiency. This gets underpinned by the current economic climate in combination with a worker shortage (although this shortage might also be self-inflicted by businesses). 

In general, AI has become an important ingredient in customer service and, more generally, CX solutions. The main use cases are ticket routing and assisting in solution search. While there are a number of chatbot solutions that base on an underlying AI, their performance often is not good enough to be really successful. Another challenge is that implementations regularly take considerable time. Therefore, the pressure on customer service departments through case deflection does not decrease enough.

My analysis and pov

As Cristina Fonseca, VP Product of Zendesk and co-founder and CEO of Cleverly, the company that got acquired by Zendesk in 2021 and that is behind this feature set explained to me, Zendesk was behind the competition. What Zendesk now does is automating the mundane tasks, therefore effectively relieving agents from time consuming tasks. The second part is that the solution comes with pre-trained models that make it ready to run in very short times. Doing the hard work and analysing that 80 percent of inquiries have the same intent across industries has been the key for this. Still, the built-in feedback loop that allows for constant company specific readjustment of the model, is important as there normally is no one-size-fits-it-all solution and there is a high likelihood for service requests changing in nature over time. This is also true for the revenue drivers prioritization of requests, which is likely to be at least industry-, if not business model and therefore company specific. I am curious to see how this works across industries.

Making these features available as part of the suite and not as an add-on is a smart move as it doesn’t change the pricing for existing enterprise edition customers and potentially drives more customers to choose this edition.

With this solution Zendesk gains an edge, albeit probably of temporary nature. Clearly, the company should continue on this path to maintain it, maybe adding coaching capabilities that help agents and supervisors to train themselves and each other. Additionally, if the solution works as expected, why not also using it for case deflection. After all, from a system perspective, it does not really matter whether an agent is guided on how to best resolve an issue or whether the customer him-/herself is guided. From a company perspective, it does. Fonseca maintains that the market is not yet ready for it. This is probably right as many chatbot implementations are rather underwhelming. Still, coming from a working solution can change this. In addition, Zendesk should add an explainability capability, although Fonseca maintains that this is not a priority as there is too much turnover. Still, a good deal of productivity improvement and efficiency will come from achieving a stable workforce. With an explainability component Zendesk can achieve two goals: first it helps coaching the agents and second it will increase the acceptance of the system. Untrained agents will more easily accept the system recommendations, effectively treating them as prescriptions. In which case they better be accurate. The confidence score that is already provided is a first step into this direction. 

LEAVE A REPLY

Please enter your comment!
Please enter your name here