TORONTO — Most people in the car industry will associate "artificial intelligence" with obstacle-avoidance systems and similar high-tech gear. But AI is also taking on a key role on the marketing side, the Automotive News Canada Congress heard Thursday.
Whether the subject is an image, a voice or a collection of statistics, "the science of analyzing data is exactly the same," explained Foteini Agrafioti, chief science officer for RBC Royal Bank.
And by recognizing the patterns in, for example, car purchases, an AI program can predict not just when a subject will want to buy a vehicle, but what type of vehicle and even the financing package that buyer will seek. A trial analysis by Agrafioti's researchers yielded results that outperformed human predictions.
The payoff? An opportunity to create "highly personalized" offers that will benefit both buyer and seller, and a potential 70-per-cent reduction in the pool of people being marketed to – a big benefit for dealers and car companies.
AI, or machine learning, has been advancing rapidly since breakthrough research by a University of Toronto team led by Geoffrey Hinton into neural networks, also called "deep learning." Its commercial potential, however, is still in its infancy, although Agrafioti's 50-person research division at RBC has started to deliver practical analysis to customers of the financial giant.