VIDEO | Fidelity Investments Innovation Leader: I Like to Brand AI as Assisted Intelligence

(US and Canada) Brian Coogan, Innovation Leader at Fidelity Investments, speaks with Savio Rodrigues, VP Client Partnership at Trianz – USA and CDO Magazine Editorial Board Member, in a video interview about his background, the applications of AI in financial institutions, and the biggest AI myths.

Coogan initially worked for financial services companies in back office operations. It was while working with the anti-money laundering programs for banking customers at EY that he realized the power of data and the concept of data science. He describes himself as a broker between business and technology partners trying to solve problems.

Speaking about the applications of AI in financial institutions he says that Fidelity deployed its first machine learning algorithms in 2014 to eradicate bad actors internally. It was successful and was trained on large data sets collected over 75 years. Since then, the organization has built hundreds of iterations of different learning models.

Coogan elaborates that AI/ML models are not static, they have to constantly relearn and reevaluate while also calibrating for the human bias in development. He suggests that it is impossible to be perfect and data scientists need to accept it and be as close to perfection as possible.

Sharing his experience of addressing challenges at Fidelity, Coogan reveals that one of the biggest challenges for the company is taking unstructured textual data and images of documents and extracting the data from them. It incurs costs in buying data and manpower. The organization has been blending different techniques like using technology used in driverless cars for optical character recognition to automate the process. He however maintains that there are still roles for humans in the early stages of AI and terms AI as assisted intelligence.

When asked about AI myths businesses need to avoid, Coogan says that people need to understand that these technologies are not perfect like humans. But unlike humans, machines do not get offended on pointing out their mistakes and can continuously monitor and make adjustments.

The CDO Magazine community welcomes Amitabh Seli to our growing Global Editorial Board membership.