Robotic process automation (RPA) used to automate brokers’ repetitive tasks has become a much more mainstream technology over the past year or two, the co-founder of software robot tech company Quandri told Canadian Underwriter in an interview.
“A couple of years ago, it was really early adopter-type technology,” Quandri co-founder and CEO Jackson Fregeau says. “Whereas now it’s, I think, more mainstream technology and the reliability [and] accuracy… has really improved.
“Implementation of that, it’s really improved to a level where it can be adopted quite easily by a mainstream brokerage.”
One to two years ago, RPA was “still very much an unknown to brokerages,” Fregeau says. “They didn’t know how it would apply to their business, what to do about it…
“When we’re talking to brokers now…it’s, ‘Where do I apply this? I know I need to. I understand the technology, I need to use it.’ From our perspective, it’s a much more mainstream technology than it was even a year ago, but a pretty broad adoption in the market as people realize how beneficial it can be.”
Fregeau’s findings match those of brokerages in Ontario, which are rapidly embracing bots to automate repetitive tasks, the provincial brokerage association told CU in late January. Brokerages are using single and multiple bots to perform different tasks, says Norah Black, vice president of marketing and communications at Insurance Brokers Association of Ontario. Brokerages typically start with bots for daily insurance rating or customer service tasks such as renewal reviews or document management, she says.
“The one thing I’m finding is that at every single brokerage I’m in, once I deliver the first bot, they are all clamouring for more bots,” Curtis Samoy, the founder of brokerage RPA provider Clementine Tech, told CU earlier. “They want more; they have a lot of processes.”
Full control
For its part, Quandri recently launched the Control Room — a platform allowing insurance brokerages to have full control over their AI robots “without needing to write a single line of code or having any technical resource from our team or their team involved in it,” Fregeau says.
Brokerages can do things like configure the robots, test, start/stop them and move them in between production and demo. This is unique as most bots (inside and outside of insurance) typically don’t allow user control.
“You don’t need to understand how it’s built, you don’t need to understand what technology is actually yielding the output,” Fregeau says. “All you need to understand is, ‘Do I want my robot to look for this coverage?’
“It’s really just the insurance experience that you need to put in there and then the robots go and do the work. It simplifies it, making it more accessible for a really broad segment of the market.”
Vancouver-based Quandri has approximately 80 (brokerage) customers, the majority of whom are using the control room tool. Between one-half and two-thirds of these customers have multiple robots running.
Quandri works on several use-case scenarios, focusing on renewal reviews, the download process into the broker management system, and activity and document management. For example, the renewal review bot will compare last year’s policy to this year’s, looking at endorsements, extracting data out of those documents and broker management systems, and then summarizing upsell and cross-sell opportunities, missing discounts, gaps in coverage, etc.
Over time, these bots become better, more efficient and add more functionality, Fregeau reports. Using the renewal reviewer bot as an example, he says the bot used to have about 10 coverages to look for a couple of years ago, but can now extract about 100 different attributes from policies.
“They’re localized to specific regions, they’re specific to the carriers that we’re identifying as well,” Fregeau says. “So, the intelligence level of the robots and the automation has increased pretty substantially. The recommendations being made are pretty nuanced and specific to that brokerage, to that carrier, to that region.”
Time-to-value has also increased. Clients can now be up and running with the control room in a matter of weeks compared to before, when it could take up to a year.
Feature image by iStock.com/Khanchit Khirisutchalual
