If you have recently contacted a customer support center, chances are your query was answered by an AI assistant or agent. The use of AI assistants in the enterprise is a growing trend, and one that Gartner predicted will reach 25% in 2021, despite only being 2% at the time of the prediction in 2019. ResearchAndMarkets.com expected this market to reach $25.63 billion by 2025.
The COVID-19 pandemic did not slow this trend, either. On the contrary, it accelerated the adoption of virtual digital assistants for enterprise applications. But what exactly are virtual assistants, how do they differ from chatbots and why are they increasingly being used in the enterprise?
The rise of virtual assistants for enterprise applications
Virtual assistants are AI-powered conversational agents, mostly deployed in environments requiring dialogs and conversation as the medium of conducting business. As mentioned above, customer support centers are an ideal example of where such technologies would be required, as customer queries are presented and processed in natural language. And there are a wide range of industries relying on call centers, ranging from banking and financial services to healthcare and hospitality.
Natural language conversation is the preferred way of communication -- it is simple, requires no technical knowledge and it is the method we are most used to. An AI that taps into this medium successfully unlocks the door to easy customer service and is what made voice assistants like Amazon Alexa or Google Home attractive to consumers.
This article is part of
Of course, this is a complicated field -- a sentence may have multiple intents and a single word can negate the entire meaning. We may easily switch context during our discussions and return to the previous topic after a while without losing track of what we were talking about.
All of these factors can be challenging for AI; thus, one of the key requirements behind the success of virtual digital assistants is being smart enough to deal with these cases without getting confused or forgetting the context of the discussion.
"Virtual assistants are equipped with deep learning and conversational capabilities that make them understand context, user preferences and sentiments. They can learn from past data and make sense of various data points to make informed decisions," said Amit Aghara, CTO of solutions at Kore.ai, an enterprise virtual assistant platform. "For example, they can clearly differentiate the meaning of the word 'visa' in two sentences -- 'Do I need a visa to travel Singapore?' and 'How much credit left in my Visa card?'"
Besides becoming smarter, another driving factor is the cost, as it is less expensive to deploy virtual digital assistants than hire and train operators. Volatility is another driving factor. "Customer demands fluctuate during a particular shopping or travel season or during unforeseen circumstances like the COVID-19 pandemic," said Jonathan Crane, chief commercial officer at IPsoft, which develops conversational AI.
Some try to respond to the sudden spike by hiring more agents, but that is costly as well as inefficient. "These jobs are also likely to be temporary, as customer centers drop unneeded employees when call volume declines, decreasing morale among those who were lucky enough to stay," Crane continued. "This could partially explain why call centers have such high turnover."
With virtual agents, these problems do not exist, as the AI can take a part of the surge in calls and let the current operators focus on the more critical tasks.
The third driving factor of virtual digital assistants for enterprise applications is the pandemic itself. Many call centers had to lay off part of their staff, yet requests for customer support were increasing significantly. The need to solve this problem pushed some companies or organizations to try virtual assistants for the first time -- and now, having enjoyed the experience, they are unlikely to go back.
Virtual assistant applications
Customer support was one of the primary targets for adopting virtual agents. Of course, not all industries were affected in the same way.
"There's a lot of interest from sectors such as healthcare and financial services," said Jen Snell, vice president of product strategy and marketing at Verint Systems, which develops intelligent virtual assistants. "These are two areas where customers naturally have a lot of questions that are both urgent and important." Other experts also highlighted the two, with retail coming in third.
Several industries are also witnessing fewer consumer calls as a result of the pandemic, including travel, airlines, media and hospitality. However, they are expected to pick up as businesses start to reopen.
Besides customer support, virtual digital assistants are used in several other ways. For instance, Kore.ai utilizes them to assist the customer support operators by providing extra information regarding the caller. Another use case is in HR, where virtual assistants help the employees by answering their queries or informing them about their salaries.
In terms of using a conversational platform to communicate, virtual digital assistants have a lot in common with chatbots. There was a lot of hype around chatbots and they were expected to replace apps, particularly around 2016. However, the technology fell short. Many chatbot engines used simple rule-based algorithms that relied on customers asking the right questions; otherwise, they would fail. The lack of a proper AI also forced the developers to mostly rely on input buttons, almost removing the conversational element entirely.
Part of why we are now experiencing virtual digital assistants for enterprise applications is that companies want to distance themselves from chatbots. The other part can be attributed to the increased integration of virtual assistants with a company's existing technologies, allowing the system to go beyond simple question-answer format and fulfill tasks.
Chatbots could also be powered with a proper AI, but virtual assistants are marking a new age -- one that wants to differ from its past.