Per Ottosson is the CEO at Artificial Solutions, a leading company in Conversational AI.
At the Mobile World Congress in Barcelona at the beginning of March, two conversational topics kept coming up: 5G and the metaverse. When I say metaverse, I’m not just talking about “Meta,” Facebook’s recent rebrand, but about the many different virtual worlds that companies are creating.
The metaverse has been around for a long time. Just look at what Microsoft has been building over the past few years. But the announcement that Facebook would be allocating billions into its own version has clearly shaken the industry and re-energized thinking around the concept.
But will it really take off? I believe there are a few reasons why consumers and brands have been slow to warm to the technology and why it is not a ubiquitous part of our lives already—one of which is the inadequate quality of voice and chat assistants.
And Meta knows it.
In a recent two-hour broadcast that began with a high-level introduction from Mark Zuckerberg, Meta’s Conversational AI Tech Lead, Alborz Geramifard, said: “Assistants today—whether via voice or chat—are generally underwhelming. We’re sharing the challenges developers and engineers face when attempting to build useful assistants and how we can navigate these challenges as we build for the metaverse.”
He’s not wrong. The conversational customer experiences that people have when engaging with chatbots, virtual assistants or voicebots is still generally bad. I should know, I work for a company that has been at the forefront of this sector for the past 20 years.
Meta is investing a huge amount of resources into this area, and Zuckerberg seems to want to build a personal assistant that betters the likes of Siri, Alexa and Google.
But it’s really not that easy.
Conversational AI (CAI) is wildly complex, including ethical challenges that require serious consideration, and creating technology from scratch that can deliver truly intelligent “human-like” responses will be a challenge, and that’s exactly why voice and chatbots are generally “underwhelming.”
Meta knows it must have the capability to integrate best-in-class CAI for its metaverse to work. During the same video broadcast, Mr. Geramifard went on to talk about “Project CAIRaoke,” describing it as: “breakthrough research that aims to make assistants more helpful and interactions with them more enjoyable. It’s an AI model created for conversational agents. It works end-to-end, combining the four existing models typically used by today’s assistants into a single, more efficient and flexible model.”
The Metaverse’s CAI Challenge
Here’s an example of why the metaverse may fall down without CAI technology.
Imagine you find yourself visiting the Nike virtual store to make a purchase in the metaverse. Your avatar can try on all the different shoes on offer, you can hold up different shirts and literally drop them into a virtual basket, ready for checkout. But suddenly, you realize you’ve picked out the wrong sizes and that your nephew doesn’t, in fact, need an XXL set of joggers and you don’t know where the kids’ section is.
So, what do you do?
In a physical store, you would walk up to the closest assistant and simply ask a question. Herein lies the issue. Without anybody to help in a virtual environment, users could be confused and the experience would end badly when they have a question or problem.
This is where the metaverse could suffer. Every virtual brand store and every virtual space that real people occupy inside the metaverse will require virtual assistants to point, guide and support people in their journey.
Do you really think Nike would want to employ thousands of virtual shop assistants? No.
The answer is conversational customer experiences.
With an inside view of the technology at the cutting edge of the CAI sector in my role as CEO in the space, I will be paying extremely close attention to how this develops over the next year, and I think everyone in the industry should as well.