A customer’s loyalty to a brand is often set by a series of interactions—on- or offline—that tie customers to the company. Yet consumers can be fickle. Customers are increasingly discerning and expect “on-demand” personalized experiences and excellent service across all channels of interaction in exchange for their loyalty. In fact, recent studies show that 59% of consumers say they would leave a brand after one negative experience.
Because of high consumer expectations, companies must cultivate and reinforce brand loyalty with each customer interaction. Over the past several years, contact centers have played a much larger role in maintaining that loyalty. When social distancing prohibited most in-person contact with businesses, contact centers were no longer just another touchpoint in a customer journey; they were often the most important one. Zoom is one company that is bringing AI to contact centers to support that critical touchpoint.
AI models have gotten smarter, and generative AI has made its way into tools that people interact with every day. We’ve seen that in integrations into Microsoft’s Bing search engine and 365 Copilot, as well as Google’s Workspace as a productivity assistant, to name a couple. Now, Zoom’s Virtual Agent is bringing human-language chatbots to call centers. Let’s take a look at how generative AI chatbots are becoming critical to customer satisfaction and what it takes for these bots to succeed. We’ll also uncover how Zoom developed the bot that “Zoomies” (Zoom employees) call “Zoe.”
Zooming in on the need for better customer experience
At the beginning of 2020, when Covid-19 took the world by storm, bank branches closed, a huge amount of shopping moved online and many people turned to Zoom for their personal interactions. In my family, like many others, everyone was on several Zoom calls a day. That included business meetings, online school, family gatherings and even wine tastings. For many, these meetings were a natural extension of their now-remote workday, and they were pretty easy to navigate. For others, the learning curve was a bit steeper.
With all of this sudden—and often new—interaction, Zoom experienced a significant surge in growth. In the first fiscal quarter impacted by the pandemic, Zoom saw a 169% increase in revenue; in the subsequent quarter, that number skyrocketed to 355%. New customer subscriptions accounted for 81% of revenue. All of that was great for the company, but the sudden surge created a new problem to solve: customer service.
During the early going of the pandemic, companies scrambled to put chatbots in place to help manage customer issues. Companies like Zoom that already had chatbots up and running found that the sheer volume and complexity of interactions were more than they were prepared for. Before injecting conversational generative AI into chatbots, interactions with customer service bots were driven mainly by decision trees that often kept customers in a frustrating continual loop and away from reaching a human. If customers did finally encounter a live customer service rep, the rep wouldn’t even have easy access to historical data on the account or the context of the customer’s inquiry.
Zoom recognized these limitations and set out to fix the company’s customer service user experience. What happened next is a surprising success story: the chatbot built for Zoom’s customer support was so successful that the company turned it into a product offering—Zoom Virtual Agent.
The pain point that birthed a bot
Once Zoom started releasing new features for its customer service chatbot to accommodate a growing and more diverse user base, the company logged millions of customer requests and chatbot sessions monthly. Zoom used that data to solve one of the company’s growing pains by creating the intelligent, scalable Zoe.
Out of the gate, Zoe delivered an impressive self-service rate of 93%, meaning that 93 out of 100 customer inquiries were effectively resolved without requiring assistance from a live agent. This resulted in a staggering monthly cost reduction of more than $13 million.
Bot + human is greater than the sum of the parts for customer service
For businesses, it’s obvious that chatbots help reduce the workload on contact center reps, which can also mitigate staffing issues. After all, chatbots work 24/7 and are easily deployable and scalable. This is particularly critical in incident response, for example if there’s a mass travel interruption from a big storm. Moreover, a bot’s ability to quickly resolve most issues frees up precious human interaction time for the trickier problems—or for customers who just really want to speak to someone live.
An independent survey commissioned by Zoom showed that when consumers access customer support, most want a quick, knowledgeable and friendly resolution. Consumers are less concerned about the channel. This is why brands increasingly steer people toward efficient and user-friendly digital interfaces for assisting customers, reducing the workload on contact center representatives.
The value of knowledgeable “agents” in building customer trust and loyalty
Zoom’s continuous training and development for its human agents to ensure they have up-to-date product knowledge and expertise has helped earn the company’s best-of-breed customer satisfaction scores. The Virtual Agent, based on an AI large language model, is essentially no different, because it is also trained on brand- and product-specific information. The chatbot can answer inquiries correctly because it is trained very specifically on company data, service manuals, product sheets and more.
In addition, when a chat is escalated, roughly 70% of customers anticipate that support teams possess comprehensive information about the reason behind their phone calls. The Zoom Virtual Agent's intelligent routing facilitates the escalation of a chat to a live agent, in part by transferring the chat history—and therefore the context of the specific problem—without requiring customers to repeat themselves. This simplifies the interaction between the customer and agent at the same time it equips reps to resolve issues faster and more effectively.
Adding Virtual Agent to the Zoom portfolio
Zoom’s strategy of expanding its offerings beyond virtual meetings is very smart. Nearly everyone has used Zoom and is at least mildly familiar with the interface. By incorporating other points of connection for people—from Zoom Events to Zoom Omnichannel Contact Center—the company is using familiarity with and affinity towards its brand to promote solutions well beyond its meeting capabilities.
Virtual Agent integrates with contact-center-as-a-service (CCaaS) and CRM software and connects seamlessly with Zoom Contact Center as well as leading CRM providers like Zendesk and ServiceNow. An ecosystem approach is wise for Zoom; it brings familiar experiences to new deployments, but also makes room for the reality that companies have mixed tech stacks.
Finally, Zoom is helping address the painful fact that call centers are notoriously high in employee attrition. By easing some of the volume and stress on reps and helping them focus on the moments that matter, Zoom fosters a scenario in which its enterprise customers can deliver exceptional customer service while also improving the employee experience.
Chatbots are becoming so advanced and human-like that sometimes it is hard to identify who (or what) you are chatting with. An intelligent chatbot like Zoe also produces advanced real-time analytics to help improve products and inform future interactions. While I do not believe that contact centers will get rid of human agents anytime soon, chatbots can drastically increase customer insights, reduce repetitive work, mitigate negative interactions arising from customer frustration and in general help the humans in contact centers do their jobs better. For all these reasons, Zoe deserves a round of applause.