Datalog.ai Launches MyPolly, an AI Platform for Continuous Conversation
Artificial intelligence start-up datalog.ai today announced its MyPolly platform, which aims to allow continuous conversation with virtual assistants by enabling the assistants to interpret human input. Over time, the platform allows virtual assistants to revise their responses based on the information they receives from users, with the goal of powering a dynamic interaction between human and machine.
"What was broken—and this is true in bots and virtual assistants, and in virtual assistants we include Amazon Echo with Alexa, we include the new Google Home, any consumer device where you speak to interact with it—the problem today is, they're set up like search engines, so instead of typing, we speak into them," says Jack Crawford, founder and CEO of datalog.ai. "They do great speech recognition, but they can't take you to the next step, which is to have a conversation. This is what we see that you'd want to do with any type of robotic consumer device.
"You want to have an exchange with it; otherwise, you just go and type. Same thing is true with Siri—it just saves you that manual typing, which is a huge benefit, but there's nothing beyond that. With our product [you can] carry on this conversation with the machine, and that makes it more humanlike."
In a separate statement, Crawford illustrated this point using the example of a pet's name. A user could tell MyPolly the name of his dog, and the platform would recall the dog's name at a later point in the conversation. In this way, MyPolly becomes more personalized over time, with the goal of enabling the user to have an ongoing conversation with the technology using the solution.
"This is the first component in the platform. It is very advanced, but we're going to the next level where we can bring this technology into internet devices in the home, in offices, or in an industrial setting, so that not only are you speaking with the machine, the machine is speaking with other machines and collaborating to give a message," Crawford says. "[We're looking at] IoT devices and [giving them] the ability to speak with each other and have their conversation be a collaborative response."
He likens this process to walking into a conference room where there are five people, asking a question of them, and, based on their talking among themselves, receiving a unified response as opposed to five individual answers.