RBC's push into AI uncovers Chipotle's 'Worst Queso Scenario'

The bank remains one of the few financial institutions in North America that push for AI in fintech

RBC's push into AI uncovers Chipotle's 'Worst Queso Scenario'
When Chipotle Mexican Grill Inc.’s queso was trashed on Twitter after its debut last year, RBC Capital Markets assigned a machine to weigh the effect of the social media storm on the beleaguered burrito chain.

The result was a research note subtitled “Worst queso scenario?” that used analysis of tweets and Google searches to help analyst David Palmer justify cutting estimates on Chipotle’s earnings and sales. Artificial intelligence, or AI, helped show that negative sentiments on the queso launch affected Chipotle’s brand.

Royal Bank of Canada is one of the few North American banks to have incorporated AI into capital-markets research, reshaping how analysts work and offering a signpost for where other firms are heading. RBC’s research group has six data scientists and specialists split between Toronto and New York in its RBC Elements unit.

“It’s a critical must-do for any research department," Marc Harris, RBC’s head of U.S. research, told Bloomberg in a recent interview. “Giving great sell-side analysts access to great data-science resources is going to be mission critical to their success – not 10 years from now, but over the course of the next three to five years.”

Read more: TD Bank invests further in artificial intelligence

The purview of artificial intelligence now includes machine learning, which is the ability for computers to learn by ingesting data and processing text. The financial industry is rushing to embrace the nascent technology to help in areas including trading, investing, and automating interactions with customers. Hedge funds and money managers also have adopted AI as a research tool.

Europe has seen some AI innovation for equity research due to MiFID II regulations, according to Jodie Wallis, managing director for artificial intelligence in Canada with consulting firm Accenture Plc. In the U.S., firms like Goldman Sachs Group Inc., Morgan Stanley, and Bank of America Corp. have been “quite vocal” about using such technology, he said.

“Most large banks, and probably all of the big banks in Canada, are at least piloting AI in equity research, but with varying degrees of commercial success at this point,” Wallis said in an interview. “There are a couple that are ahead in terms of their ability to extract real value out of those investments now, but that is definitely an area where all the banks are looking.”

Royal Bank Chief Executive Officer David McKay has publicly embraced AI since 2016, bringing on board industry and academic experts and establishing labs and partnerships. RBC has more than 200 data scientists working in various departments including capital markets. Advances in computer processing power have made what was once the realm of science fiction now achievable for everyday operations.

“It’s not easy to do this, but the rewards of doing it right are definitely proving to be worth it,” RBC’s Harris said. “This is a long-term marathon, not a short-term sprint.”

In the Chipotle note, data showed that negative tweets outnumbered positive ones in the week after the queso launch, and that sentiment remained negative, though improving, for some time. Two years ago, finding correlations between social media data and stocks didn’t work, but advances in natural language processing with AI makes this data more valuable in finding impact, according to Harris.

Harris, who describes using AI in financial research as being “a first- to second-inning phenomena," said he expects more work like that done on the Chipotle note.

"We’ve now got literally several dozen projects in our pipeline that follow a similar track,” Harris said. “We’re basically trying to look at energy, industrials, a lot of different verticals.”

Related stories:
RBC pilot-launches AI-powered services offering financial insights, savings solutions
Thought leaders betting on analytics, automation, and big data