Russell Ward, president of Toronto-based chatbot maker Massively.
When I listen to an automated answering system and its recorded voice, I also hear another voice.
It's my father's, inside my head, and he's saying, "Just get me a real person."
The tech world knows all about this reluctance in some, even as countless other users accept the speedy, semi-personalized automation being dished out by algorithms and machine learning.
That preference for a human being could even be seen as driving the evolution of chatbots and the sector's rapid expansion.
"There are two ways to think about that: If there's a question the bot can't handle, we want to make it as quick and as seamless as possible for a real person to take over that conversation," said Russell Ward, president of the Toronto-based chatbot maker Massively. "The other function is what we would refer to as triage-ing, where the bot can get you to the right person quicker."
So, if you're calling, say, your cable company, the customer-service bot would get you to the right technical helper more quickly. But there are other nuances, too, that chatbot makers are going for.
Massively, which is looking to grow its presence in the United States, started in January, 2014, as a division within House of Cool, a pre-production studio in Toronto, which specializes in story-boarding and other work involved before a movie begins shooting, particularly with animated films. The thinking behind the creation of Massively was this: Wouldn't it be neat to have the characters talk to you beforehand?
This led to the company's niche in creating chatbots focused more on engaging users for longer periods of time, rather than just providing fast answers (even though speed and accuracy remain key in order to hold users' fickle attention spans). The division was spun off and became its own independent company in December last year.
Emulating a live person, though, isn't really a chatbot's purpose. Sure, the automated dialogue, typically on one's smart phone, has the pretense of conversation (such as the text-based recipe suggestions on Facebook Messenger from the Philadelphia Cream Cheese bot). But chatbots are evolving away from simply trying to mimic human conversation.
"The technology still has to come a considerable ways. That being said, the advancements are coming fast and furious," Mr. Ward said. "All the big guys, quite honestly, have invested a lot and have, in some cases, staked their futures on conversation and artificial intelligence powering those transactions."
Competition is as varied as the technology. If you talk to 20 different chatbot makers about their technology, you will get 20 different answers, Mr. Ward said. "Businesses are being built on their ability to make smarter recommendations, or better recommendations, and they are selling those algorithms as a service."
And it turns out that, perhaps ironically, a live presence does matter, at least in drumming up business. Massively has a representative in New York, basically project co-ordinator-slash-salesrep, who is paid by Massively for the amount of work brought to the company. Massively is looking at a rep in Los Angeles, as well.
The work is still done entirely in Toronto, but there is a need for physical representatives abroad, Mr. Ward believes. "Our sense is that if we lose a piece of business, one of the reasons is because we're not physically present there."
Massively's niche isn't so much to answer customers' questions as quickly as possible, but to make chatbots that present bite-sized entertainment experiences, often in the guise of a conversation on their smart phones.
"How we try to differentiate ourselves is on engagement," Mr. Ward said. "So, we're really focused on building experiences that people enjoy and where they want to talk. In most cases, these are marketing applications. … We build experiences that are more interesting, that are fun, that are gameified."
The company's chatbot Epic Reads campaign includes a marketing project for publisher HarperCollins Publishers launched on Facebook Messenger and the messaging app Kik that delivers an emoji-accented discourse on book recommendations. Its Philadelphia Cream Cheese campaign features recipes and is laced with emojis, all the better to gauge the whims of a respondent conversing by text and to better move the conversation toward his or her individual tastes.
The company recently hired a linguist to help link the technical side of the business (which looks for language patterns in users' responses) and its creative employees (that is, those who script the dialogue).
"When we talk about language understanding, they [chatbots] don't necessarily understand intent, and they don't understand context very well. And so, that's where the linguist comes in and works in that space, in order to bridge the conversational side and the technical side," Mr. Ward said.
Understanding voice commands is particularly a growing area for chatbot makers. "If I'm in my kitchen, I can just talk to my Amazon Alexa or Google Home to get information I need. So, there's a level of convenience that we're trying to deliver," Mr. Ward said. (By recognizing voice commands, Amazon Alexa and Google Home perform various digital tasks, such as playing music or turning on a home appliance, and are major, newer platforms for chatbot makers.)
But still, technology has to meet users' rising expectations. "The consumers or the end-users of services are becoming smarter and smarter, and their expectations are becoming higher and higher," noted Hussam Ayyad, senior director of programs and partnerships at the startup institute DMZ at Ryerson University in Toronto.
"You're selling them experiences. You're not selling them services. There is a very big difference," Mr. Ayyad explained. A service is, for instance, a simple cab ride. An experience is a ride-sharing service, in which you can rate the service, pre-pay, track the location of the vehicle online, etc., he said. "Experience is a by-product of a service. The by-product is the value that the consumer perceives and receives from that service."
The emphasis, then, in order to improve the experience, is on the nuances, on understanding what users really need – needs that might fall through the cracks of automated systems. "The connection between the words they said and the context of those words, that's where the true innovation is happening now," said Massively's Mr. Ward.