New initiatives are never easy, outcomes never certain. When CIBC opened its lab at the Communitech Data Hub in Waterloo in May of 2017, the underlying hunch was that the lab would add value, but it was impossible to know to what degree or if the hunch was even on target.
Eighteen months later, the picture is clearer.
“It’s been an incredible success story,” says Santanu Pal, CIBC’s Vice-President of Enterprise Advanced Analytics.
Leveraging success is why CIBC tripled its Data Hub footprint last month to 1,100 square feet and why, by the end of this year, it will have more than doubled its staff, to 15. It has additionally doubled the number of co-op positions it’s offering from four to eight.
“The reason we expanded was, frankly, because of the success we were having,” says Pal.
Big banks, it would surprise no one to learn, are models of stability and predictability and, additionally, heavily regulated – government-mandated hallmarks which keep the nation’s commerce and money flowing smoothly and predictably.
The challenge for Pal, and lab director Bart Piwowar, was to find ways to build solutions for the bank and quickly change the status quo within a heavily regulated framework – and to do so in an organization with more than 40,000 employees.
The result?
“We actually do things [at the lab] in four to six weeks that banks typically take years to get done,” says Pal. “What we can do very quickly is we can build prototypes and showcase their value very quickly to the bank.”
The projects, Piwowar says, have focused on analytics – using data and artificial intelligence to help the bank understand its customers and to hear what its clients are saying.
(Communitech photo: Sara Jalali)
“[CIBC does] a lot surveys online,” Piwowar explains. “We survey our clients, we ask them to provide feedback – a free-form text response. What are our clients are telling us? More specifically, what are their top pain points? What are the top positive things we’re doing?
Answering those questions, and doing so at scale, was the problem.
“If we get 10 [survey] responses back from our clients, humans are capable of understanding that,” Piwowar says. “If we’re receiving 4,000 [responses] a month, that becomes a challenge from a human effort standpoint.”
The solution was to use AI and natural language processing to sift, sort and make sense of the trove of survey responses and the accompanying data.
“We developed a tool, a data product, that analyzes these responses as we get them from the field,” says Piwowar, generating a “net-new capability that wasn’t there before.”
Another project, also utilizing AI, centred on the bank’s back office, where teams do manual correction of documents and information. The work is time and labour intensive, so the data lab team produced an AI process to “assist the humans,” turning an activity that once took 20 minutes into one that takes two.
“So the AI is not doing the work of the human, but it’s assisting the human to be faster and more optimal,” says Piwowar. “The human can do the more complex, cognitive work and the AI piece takes care of the less complex, manual work.”
Pal frames the process like this: “There are certain things humans are not good at. There are certain things technology is not good at. So how do we spin it in a way that gets the best of technology and the best of humans and create an experience that is really differentiated for our clients?
“Data analytics have a huge role in pulling it all together, in connecting the dots, and saying this is the client’s need and what the clients are looking for. Data analytics provides that connective tissue.”
A second mission assigned to the lab when it opened was talent generation – “to attract the best and the brightest from the STEM programs you have at University of Waterloo,” says Pal.
The hope was that the lab would become a hands-on way to showcase the bank to graduating students as a place to work on interesting problems and to show that banks have evolved beyond their old stereotypes.
Once again, Pal is thrilled at the results. He says that when CIBC opened up submissions for resumes for its first co-op term it received 16 applications for four positions. Its latest co-op has, he says, attracted “150 to 180 applicants to the same four positions.
“We’ve been able to attract talent. We’ve been able to grow the talent pool. We’ve made this a place where people would like to work.”
The final objective CIBC had when it moved in was to engage with Waterloo Region’s innovation ecosystem. Once again, the outcomes tell a tale: collaborations have unfolded with startups SkyWatch and Kiite, which are also Data Hub tenants, and Thomson Reuters, which operates a lab at the Tannery building in Kitchener.
“I’m very happy about it,” says Pal. “I think the pace at which we’ve been able to do some of this stuff has been surprising to me. It’s been a liberating experience for everyone.”
Big data, big bank, big win.
