Webinars / Blog Home

Where AI is Interacting with Customers in Banking

By August 27, 2020

  • facebook
  • twitter
  • pinterest
  • linkedin

During this webinar, industry experts discussed the main areas where AI is interacting with customers in banking. We’ve included a short transcription of the webinar, beginning at 4:17 of the webinar.

Dimitri Bianco, Santander Bank: The first question is going to be, what are the main areas where AI is interacting with customers? Michelle, do you want to take this for us to start?

Michelle Bonat, JPMorgan Chase: Sure. Thanks, Dimitri. So this is my own perspective as an employee at JP Morgan Chase. So let me lay the blanket for where I get that perspective. JP Morgan Chase is the country's largest bank by revenue. We have a great opportunity to tackle really important opportunities and challenges related to AI, both in the resources and the amount of data. So we work with twelve billion tech spend a year and have access to enormous amounts of data with relationships with about half of the US households, processing six trillion dollars in payments every day. And in terms of the digital world, which I live in, the benefit of 53 million digitally active customers and 38 million mobile active customers. So, from the perspective of a technologist, I have great tools and resources to work with. 

In terms of where the main areas where AI is interacting with customers I’ll speak on some things I see our firm doing and some things I see in the industry more broadly and some things that I'm particularly really excited about. JP Morgan Chase already in production for AI are things like fraud detection, trade execution, customized credit card offers, and many others, as are other institutions working in that. When I think about it, I read something that captures it really well, so it’s core, adjacent to core, and transformational as the way AI is being worked into banking, or more broadly financial services. 

So core at our firm, that's banking, investing, credit cards, payment services, and then adjacent, I think of it as new areas or developing new operating models using AI to enhance the experience. With those new areas, I think of it as call center. So there's work being done in call center to either categorize the tickets better or AI chats like that and then from a transformational perspective, that really goes even beyond banking into anything that is possible. So that's really everything else. 

Personally, I'm really excited about the uptake in NLP, natural language processing, in banking. Previously, we all came up mostly looking at numeric data, but as is often said, the non-numeric data comprises likely 80% of all the data out there that we could use and so 80% has been largely untapped. As banks and financial institutions get more experience, more tools to leverage that unstructured data with NLP, I see that as a really exciting way to grow the efforts.

Dimitri Bianco, Santander Bank: Anyone else want to jump in on this question here?

Mary Miras, Goldman Sachs: This is Mary and I would just add that I think Michelle's point about NLP is really well taken. I think that's one of the things obviously that AI in any sort of machine learning is most depending on is data. She pointed out that there's just this huge amount of data to date that’s untapped and I think that potential there is really incredible.

Dimitri Bianco, Santander Bank: Have you seen any issues with processing NLP? I know from a risk perspective, we’re always concerned with data quality, customer interaction (do you have good retention with that). Have you guys seen similar issues or do you guys have more of a clean set of data to work with?

Michelle Bonat, JPMorgan Chase: With NLP, the reality is if it was easy everybody would do it and it's not easy. So, data is data, text is text, everyone has the same challenges and we always think it won't be 80% of the time doing data engineering and clean up and then it always is. So, for financial institutions collectively we all have those bigger layers of regulation which certainly we all adhere to. So the inherent challenges and opportunities are that NLP does have unstructured data to look at, but then layer in all the ways that you get that working and all the possibilities to do that and then leverage it into deep learning and transfer learning and reinforcement learning and all the cooling bells and whistles with the the regulatory wrapper. So that is certainly the challenge but certainly the opportunity as well. 

Dimitri Bianco, Santander Bank: I think it's a great point because I think for a lot of people on the outside looking inside of banking, they always think we're very slow and behind the curve, but I think other industries like tech where you have more excitement, there's a lot less regulation. For banks, we’re definitely taking more of that cautious perspective on making sure that we're doing it correctly and safely for our customers.


Learn more and watch the full video on YouTube:

Recent Posts