During our inaugural What’s Next seminar, experts weighed in on the leading trends and developments in the Conversational AI market.
It’s a daunting prospect to follow every new trend and development in the Conversational AI landscape, considering the lightning speed at which the market is moving. It’s why we focused our inaugural What’s Next hybrid seminar session on Conversational AI — to provide clarity and thought leadership for businesses interested in advancing their operations with this transformative technology.
(Listen to a replay of our most recent seminar, What’s Next in Intelligent Automation, by clicking here!)
What’s Next in Conversational AI featured new insights from Amelia CEO Chetan Dube and Anthony J. Bradley, Group Vice President of Emerging Trends and Technology Research at Gartner (the replay is available here). Although automating password reset requests with Digital Employees continues to be revolutionary for many organizations, looking ahead, Chetan argued that we should aim to further bridge the gap between what humans can do and what Digital Employees can offer.
When companies receive calls, emails and messages from customers, their requests are often “problem cases,” Chetan described. For Digital Employees to provide support that is equivalent to what humans can offer, we must close the intelligence gap between human and digital employees. “We have to cross the chasm from classification, using DNN (deep neural network) models… to comprehension, to understanding,” he said.
Chetan offered three foundational shifts that need to occur in order to move beyond the novice capabilities of virtual assistants to the more advanced realm of intelligent virtual agents (IVAs). First, instead of teaching AI, which is a manual act, AI must learn automatically. Second, AI must learn from the copious amounts of “digital exhaust” that exists within organizations. “We really have to develop a compiler for natural language,” said Chetan. “We have to automatically glean that imbedded intelligence.”
Finally, in order to advance conversational intelligence, we must “raise the bar of IQ on our (virtual) agents,” said Chetan. “We’re at the tipping point. If we are able to push the frontiers of IQ… we can actually come closer to that World Economic Forum hybrid society realization where by 2025, 52% of employees in your organizations are going to be digital.”
So that begs the question: What will virtual agents be capable of doing within the next twenty years?
Anthony led a team of three analysts in a yearlong study to explore this question. They interviewed 150 technology providers, and 300 leading-edge technology adopters within the virtual agent space. His team found that we are at the start of the “advanced virtual assistants” era. He shared the study’s main takeaways during the seminar.
Virtual Assistants Everywhere (VxA)
According to Anthony’s team’s research, we can expect advanced virtual assistants to overcome the obstacle of language, enabling them to fill advisory and intervention roles within the workforce. While we often think of conversations in terms of languages such as English, Japanese and Russian, Anthony said, “language is far more sophisticated and nested than that.”
There are multiple layers of terminology within each language, such as medical terminology — which is challenging for most human beings, let alone for virtual assistants. “That language barrier has kept Conversational AI and agents [at a certain level of] customer support — why? Because they’re talking to normal people and [in] normal language,” Anthony said.
Now, with models that look at sentence structures instead of just words, we can push the envelope of what virtual assistants are capable of handling. “What it’s allowing us to do is consume languages and Natural Language Understanding far more rapidly than we used to,” he explained.
Advanced virtual assistants that can understand, and act upon, niche and industry-specific terminology will add value to organizations by empowering human employees across all industries and job functions with their own personal digital colleague. In fact, Gartner predicts that, “by 2025, advanced VAs [virtual assistants] will provide advisory and intervention roles for 30% of knowledge workers.”
Multimodal: Talking to Everything
Humans currently interact with technology in relatively limited ways. Unlike human-to-human interactions, which involve interacting with our surrounding environments as well as with one another, interactions with Internet of Things (IoT) and “edge AI” technologies are almost exclusively conducted via screen.
Looking into the future of advanced virtual agents, Gartner sees the impending growth of composite AI, which enables virtual assistants to act, instead of simply conducting conversations. “Composite AI is required for really robust virtual agents,” said Anthony. “You talk to the thing, but that thing still has imbedded intelligence that it can react to, so not just understand but act, and then report back.”
In terms of impact on the Future of Work, some jobs will be lost to technology, but we can expect to gain far more jobs and entirely new industries as advanced virtual assistants will enable us to create them at “hyper-human scales,” Anthony said. “There will be lots of new jobs and lots of new industries. A lot of it will come from the fact that we are extending human capability by being able to interact with machines in new ways.”
Intelligent Composable Business
Today’s enterprises consist of several functions, departments, tools and more. While this may help organize internal structures, the reality is that many companies expect their customers to navigate their organizations’ complex and disconnected silos themselves whenever they need help or information.
Even when customers ask a human agent for help, human latency slows down overall processes. “There’s no human being in the world that is going to be able to navigate [the fixed, functional silos] in real time to get me what I need,” Anthony said. As customers, “we want a business that looks like it’s built just for us.”
So how do we get there?
“We talk about this decomposing actual business functions into some things that can be swapped around or orchestrated into something that can […] deliver you a product faster,” Anthony explained. By decomposing and re-composing business functions, companies can hyper-personalize their business. However, Anthony also stated that “without AI, there is no such thing as a composable business. You need AI to help you compose those components into something that is worthwhile.”
Unlike human agents, advanced virtual agents are capable of instantly orchestrating across an entire enterprise. As Anthony explained, virtual assistants must be able to “make decisions, access the legacy systems and make it happen on my behalf.” Some organizations are taking these advancements a step further by creating “networks of virtual assistants,” with each one being an expert within their role. “You want all the experts to come in the room and work together,” he said.