December 11, 2018 by Greg Meckbach
Using artificial intelligence, brokers can gain valuable insights from data, but making decisions based on those insights might be easier said than done.
“There is an issue that all companies struggle with. It’s one thing to gain insights out of data. It’s much harder to apply those insights into whatever your operational systems are your operational processes,” Ian Scott, partner and chief data scientist for Deloitte Canada, said of artificial intelligence in an interview.
Scott gave an example of how a broker might use AI.
“You might analyze all the customers. Using some simplistic AI tools, you may be able to identify opportunities like existing customers who may have affinity for product they don’t already have,” Scott said.
This way, AI could be used to determine that one customer is more likely than another customer to be interested in a certain product. Every customer would get a “propensity” score.
“But that’s just a number,” Scott said. “A score is not an action. An action needs to be taken – whether it’s the broker looking at a collection of propensity scores and making a judgement call – or you have the system … say, ‘Well for this customer the best offer is this product.’”
Deloitte recently surveyed consumers and businesses on AI adoption, reporting the results in Canada’s AI imperative: From predictions to prosperity, released Nov. 21.
AI adoption among Canadian firms “has remained stagnant over the last four years,” Deloitte Canada suggested.
Only 16% of business respondents said they used AI, a number that did not change from 2014, Deloitte said.
“The study found that both consumers’ and businesses’ lack of understanding, lack of trust, lack of awareness, and an inability by companies to scale small pilots were barriers limiting adoption,” Deloitte said in a press release.
Of the consumers surveyed, 86% said they do not use AI-powered tools or devices.
But that number should be higher, given that 76% of Canadians own a smartphone and likely use applications like Siri and Google Maps, Deloitte noted.
Nearly half (47%) of “early” business adopters of AI reported they were concerned about the risk of making the wrong strategic decisions based on AI.
“If done correctly, I would argue that you can have AI powered decisions that are every bit as trustworthy as if a human made them and in some cases more trustworthy. The notion that AI is a purely black box situation is a misunderstanding,” Scott told Canadian Underwriter.
Say for example you get a credit risk score from an AI algorithm. A human can understand what the score is based on, Scott suggested.
Deloitte surveyed more than 1,000 Canadians and 2,500 businesses from around the world between July and September.
In the citizen survey, Deloitte polled 1,019 Canadians from across the country. The margin of error is 3.1 percentage points, 19 times out of 20.
In the business survey, Deloitte polled 769 Canadian businesses about their uptake of emerging technologies, including AI. The margin of error is 3.6 percentage points, 19 times out of 20.
In its Global State of AI in Enterprise survey, Deloitte polled the leaders of 300 Canadian and 1,600 international large companies that are “aggressively adopting” AI. The margin of error was 5.7 percentage points, 19 times out of 20.
Respondents in the Global State of AI in the Enterprise survey had revenues of at least US$50 million and head counts of at least 500 employees.