How AI can help you find the hot sales leads

Looking for a hot lead? Why not let a machine do that for you?

Brokers and insurance companies can use artificial intelligence (AI) to cull prospects who are showing a keen interest in buying insurance products, Andrew Lo, president and CEO of Kanetix Ltd., said Wednesday at Insurance-Canada.ca’s first Technology in Action seminar in Toronto.

Insurers and brokers can, for example, identify various steps of the sales process and then train a Google machine to “find the customers we want,” Lo said. For example, customers will often visit a website several times to search for auto insurance, get quotes and then call or buy online. At the same time, Google tracks things like whether the person was searching on a phone or desktop, the operating system, language, query and location.

“So, imagine when you track all of this and feed it to the machine in addition to [information about] the conversions that you saw,” meaning customers who purchased a policy, Lo said. “That means you’ve got a great customer. What happens with the machine is, it’s going to go out and find those customers that you sold to as a leading indicator. That’s the customer you want: find more of those customers.”

What variables determine the propensity to be a customer? “The key variable is the conversion point, and you can assign which one that is,” Lo said in response to an audience question. “It could be when they got a quote. It could be that they invested in an email or they did a sale. It doesn’t always necessarily mean the sales, sometimes it’s the quote, and the sales data is secondary.”

Using a Google machine, sales for Kanetix have increased and “the leads or phone calls we send to insurance carriers are higher quality,” Lo reported. “We’ll know which carrier wants which type of customer and go help them find all of that.”

Once the right type of customer is found, Kanetix uses “conversion rate optimization” to find the “key moments of truth” or “watermarks of intent,” indicating if a customer intends to purchase a policy. To do this, Kanetix has partnered with integrate.ai to classify customers as between ‘slightly likely’ and ‘very likely’ to purchase. They feed the machine all of the conversion points and customer behaviour as they enter data on the site (such as vehicle information and driving history).

“Right off the bat, we have real-time machine intelligence looking at that data that person has input into the system,” Lo said. “By the time it gets to the quote, it knows which customers have the highest intent to call.”

Those with a high intent to purchase – for example, a Honda Civic driver who will save hundreds of dollars a year on insurance – likely don’t require a lot of marketing effort to close the deal. However, for a BMW driver, that incentive may not be as large. So, you will “want to change the user experience to get them over the top,” Lo said. “You may want to invest more in marketing because you know that person [may] convert, but if you didn’t do anything, they wouldn’t become a policyholder.”

Another way AI can be used to enhance the customer experience is to remove friction from the purchase. Voice options are one way to do that. Lo used the personal example of a Google Home smart speaker that he used for a quote. “It actually texted me the quote and a link to buy,” Lo said, adding that there is no installation or configuration required. “This removes quite a bit of friction. It only took days to build; we are working with Google to train it on our language.”