• by Chase Martin , October 2, 2016

The Internet of Things might see better conversations between humans and devices, thanks to a new competition, with agency executives saying there is opportunity for brands in this space.

Amazon recently launched a new competition for students to develop better bots for the retailer’s Alexa platform and plans to offer $2.5 million in total funding.

The Alexa Prize will sponsor up to 10 teams of university students with $100,000 research grants to develop artificial intelligence-powered ‘socialbots.’ The winning team will win $500,000.

In addition to the grand prize, Amazon says it will award the university of the winning team a $1 million research grant if the team’s socialbot ‘converses coherently and engagingly with humans on popular topics for 20 minutes.’

The general focus of the challenge is to advance the various areas of conversational artificial intelligence, such as understanding natural language, generating natural language and common sense reasoning.

Agency executives discussed the use of AI-powered chatbots last week at the MediaPost OMMA Bots and Chat at Advertising Week conference and some said they see opportunities for brands when it comes to designing these types of bots.

“We’re starting to look at brands as social beings that can essentially be thrown into any social context because these conversational platforms are inherently mobile,” said Natalie Monbiot, senior vice president and managing partner of strategic innovation at UM.

“You want an intelligent brand persona to engage with, someone that wasn’t just born today because the bot appeared today. It’s got to have all of this prior knowledge and it’s got to be a smarter, more convenient, more efficient, more fun way to communicate with a brand than ever before.”

Monbiot said that this is not unrealistic, pointing to a recent project promoting a movie release, which created a chatbot to interact with consumers as one of the main characters from the movie.

“We actually saw some staggeringly good engagement rates,” Monbiot said.

Average conversation time with the bot was 10 minutes with some of the longest conversations reaching two and a half hours, according to Monbiot.

One key in creating an effective chatbot is the data from the conversations, which needs to be mined properly for the bot to engage effectively.

To do so, machine learning is needed to create an algorithmic system that can learn more about each consumer as they interact more with the bot. This would be nearly impossible to achieve manually, especially in real-time and at scale, according to the panel.

With all of that data mining can come useful consumer insights for marketers on a broader scope, according to Pete Sena, founder and chief creative officer at Digital Surgeons.

“What I find particularly interesting is how we can understand intent at a deeper level, so how we can understand what a user’s past behaviors are and then how we can predict the future in a different way,” said Sena.

“We know that the millennial generation, while they want experiences, they want to just be able to fly through an ecosystem. They want to just tap and push and get through it.”

The Amazon Alexa Prize challenge will select the 10 teams next month to develop their socialbots through 2017.