Businesses are using micro personalization to target customers
Analyst firm Gartner reported this August that personalization has seen a 35 per cent bump in revenue
From marketing emails to suggestions of what to watch on Netflix, we've only just scratched the surface of what's possible with personalization using artificial intelligence.
"So if you've watched movie A, B and C and somebody else has watched movie A, B, C and D, you'll be recommended to watch movie D. So that's a very statistical approach to recommendations. There's a lot more we can do," said Tomi Poutanen, co-founder of Canada's Vector Institute for Artificial Intelligence and the artificial intelligence company Layer 6, which was recently acquired by TD Bank.
Analyst firm Gartner reported this August that the personalization category has seen a 35 per cent bump in revenues.
Personalization extends beyond digital realm
With micro-personalization, artificial intelligence is going beyond name and purchase behaviour.
"It's knowing where you are in that moment. Is it raining? Are you driving? Is the battery level on your phone about to die?" said Justine Melman, vice-president of marketing and communications at Flybits, an artificial intelligence personalization platform based in Toronto.
"We don't want to send you a piece of communication, like a video or something, where you can't watch it because your battery is about to die. Or we don't want to send you anything when you're driving because it's not safe. So we're going to wait," said Melman.
This kind of personalization isn't limited to the digital realm. It will cater to physical experiences, as well.
Melman said a bank could track a customer's web activity to find out what services they're looking for. Then, send a personalized message to that customer as they are near a bank branch inviting them to stop in and meet with the bank employee available at that time.
Finding a balance between exploitation and exploration
Poutanen says deep learning can bring these experiences to another level.
"Where machine learning differs from traditional artificial intelligence is that you use all of that data to inform and develop the model that's built without an analyst or a scientist trying to handcraft that model," said Poutanen.
"More importantly, they're able to take their feedback on how the customer responds to the recommendations that are put in front of them and you start to continually learn and train itself and get better."
But Poutanen says finding a balance between exploitation and exploration is necessary to ensure preferences are catered to while still exposing customers to choices.
"Exploration is one where you let the consumer explore various different points of view various different kinds of content and exploitation is one where you serve the very specific, highly tailored results," said Poutanen.
"The consumer experience is best served when there is a good balance of both."