Canadian company Voiceflow is one example of the kinds of possibilities the growing field of voice technology opens up
Voice-based assistants, such as Amazon Alexa and Google Home, are growing in popularity, to such an extent that market research company eMarketer predicts there will be 5.8 million such smart speakers in use in Canadian households this year.
So it seems only natural that a side industry would develop around this emerging technology, particularly when one considers that in the United States, 32 per cent of adults own at least one speaker.
Channeling the idea in the film Field of Dreams, “If you build it, they will come,” Toronto-based Voiceflow Inc. set out to capitalize on the technology by developing what Michael Hood, the company’s chief product officer and one of three co-founders, called “a YouTube for voice.”
The idea was to create an entertainment platform for people of all ages, consisting exclusively of voice-based content. While that offered vast potential, the company decided to begin with children’s stories that could be listened to on a smart speaker.
With the use of voice actors and audio engineers, Voiceflow released a number of interactive stories, which proved popular with parents trying to put their children to bed. But while it was the No. 1 voice app, or skill, as apps for Amazon Alexa are called, in Canada for a short period, it was still proving a tough sell.
“The problem we ran into was [the market] was growing quite quickly, but it’s tough to scale the business just because voice is in its infancy,” Mr. Hood says.
However, while in the process of developing children’s stories, the team at Voiceflow created a tool for its employees to quickly publish each individual story. Taking away some of the complicated coding that can be a hurdle for some in creating skills, the tool featured a drag-and-drop feature to allow quick assembly of interactive stories.
But the team found that publishers and authors were slow on the uptake. They then expanded their horizons and made the tool universal, allowing anyone using it to publish their own skills, whether they were stories or not.
“We thought it would maybe be cooler just to empower people to build their own [stuff] rather than try to make them play by our rules,” Mr. Hood says. “So essentially it’s like a Wordpress for voice apps, if you’re familiar with website builders.”
Since that decision in August last year, the response has been positive. Voiceflow has clients in various industries, including the film sector, which is using its tool to build interactive stories to be released in tandem with upcoming movies.
In the first month following its decision to make its time-saving tool available to other companies, Voiceflow had about 1,200 users of its tool; that number has risen to more than 3,000 now. Voiceflow has raised a total of $650,000 in funding over two pre-seed rounds. The company, which Mr. Hood said is not cashflow positive intentionally, in order to grow, currently has 12 employees, with the intention to scale to 30.
The growing numbers reflect the strong potential that some experts are predicting for voice assistants and voice-related industries in general.
The company’s future might depend on whether a deep-pocketed company such as Amazon might want to get involved, or not, cautions Alexander Wong, the Canada research chair in artificial intelligence and medical imaging and an associate professor in the department of systems design engineering at the University of Waterloo.
“As Amazon expands, they might focus on making [voice coding] easier, and so with Amazon resources it’s not a stretch for them to be able to build advanced visual tools for the users as well,” he says.
But there is a large demographic of people that a company like Voiceflow can target, he adds.
Whether they are already coders or have never coded in their life, this kind of tool allows them to have additional accessibility to the world of voice and voice-related apps and makes it easier for both sets of people to design apps.
“It’s similar to back in the old days when we had rudimentary drawing tools,” he says. “Then Photoshop came along and made it easier to create more advanced drawings, more advanced visual displays.”
For those who already code, this type of tool allows them to put their ideas into practice more quickly. “From a rapid prototype perspective, it’s actually very useful,” he adds.
Other experts are excited at the possibilities that will unfold within the voice and artificial intelligence realm. Companies like Voiceflow may be innovating in that space, but there is much more to be discovered.
“We’re only scratching the surface on how we build systems to understand language and that’s only a part of, let’s call it, the chatbot phenomenon, where we can talk to objects and they give us reasonable responses,” says Randy Goebel, professor of computing science at the University of Alberta in Edmonton.
However, he cautions there are still a lot of hurdles to overcome before this technology is worked into our daily lives. Products such as Alexa and Google Home lack depth of language understanding, so certain nuances that humans are able to pick up on is completely lost on them.
But Dr. Goebel adds that these products don’t need to understand language completely to provide a valuable service.
“The whole problem [of language] doesn’t need to be solved to create something of use,” he says.
He adds that younger generations will likely be more reliant on voice-controlled products than older ones, something he has seen first-hand on his travels.
“One of the things that’s certainly true about China is that the current generation of Chinese between the ages of, say, [18-32] … now walk around and use voice apps and voice pay [apps],” he says.
In addition, one of his former graduate students is running a startup in Beijing called Naturally, which is building a general voice app that allows one to verbally control all of the apps that are downloaded onto a smartphone.
With a market for these kinds of services, Dr. Goebel is confident that the future for voice and voice-related, AI-run technology is anything but a fad, and bodes well for companies that can provide something of specific value.
“I don’t think it will fade away,” he says. “But the high-value [offerings] are like what one of my graduate students is doing, which is targeting a specific, narrow application of languages, with which you can build incredibly accurate systems.
“You’re not trying to solve the language problem, you’re solving, if you like, using language as an interface tool and I think that’s different and that’s highly doable.”