[In the “Executive Perspectives: Machine Learning/AI” series, Expected X founder and Principal Consultant, John Sukup, interviews industry leaders on how they believe the future will be impacted by these technologies]

J: Thanks for your time today, Matt! If you could just give me a brief background on your career and how it’s led to you to the point you are right now.

M: I’ve been a digital banking transformational leader for over 20 years. I’ve worked to drive revenue and profit increases through digital banking and digital marketing by optimizing the user experience, cultivating strategic partnerships, capitalizing on value-enabling technologies, and clear product positioning. I’ve worked primarily on customer-facing platforms including online banking, mobile apps, enrollment, and the public site. I’ve been fortunate to work on several financial industry-disrupting technologies innovations and Fintech partnerships.

For the last several years I have also led digital marketing teams, optimizing customer interactions based on behavioral, transactional, social, and CRM information, and subsequently using that information to generate a “next best action.”  It’s been a dramatic change since I first started in customer experience.

I came up with a UX background, and I was interested in creating nice, simple, intuitive UI. Once site personalization became a possibility, I quickly realized a nice UI was not enough. You need to figure out, “How do I better serve these customers? How do I get the data? How do I make intelligent decisions based on that data and then proactively reach out to them again?”

When I first started in online banking, having a customer check their balance a couple of times a week was fine. Now that’s not the case – their relationships with banks are more involved and their expectations of service and security are much greater. Now we need to be inspecting their transactions. We need to be evaluating those transactions for trends. Look at their website behaviors: Are they doing something that might indicate an interest or a certain life stage or event that we can use to market them more effectively? That becomes a much more data-centric activity where AI and machine learning come into play.

J: Based on everything you’re seeing in machine learning and AI within the financial sector, are you seeing more applications in consumer-facing systems for aiding in interactions or processes or are you seeing more systems that are operations-oriented for targeted marketing or targeted prospecting. Or maybe you’re seeing an even split between the two. Could you elaborate on that?

M: I have a bias for the customer-facing systems because that’s where I’ve lived most of my career. There are certainly operations-oriented AI as you start considering information security or enrollment.  I usually think of robotics when I think of bank-end banking operations. Activities like underwriting or compliance may have several repetitive, high-frequency, error-prone, and manual steps. Robotic Process Automation (RPA) is perfect for automating those types of tasks. I had a Python developer who could quickly automate a fellow employee’s activity that historically took 4 hours—and now it takes 30 seconds. It allowed me to better utilize that employee and not backfill a similar position.

I think that a lot of individuals might not realize the wider breadth of applications for machine learning/AI systems beyond making predictions—there’s a world of opportunity just getting from manual to automated right now.

J: Financial services are one of the top industries investing in AI and machine learning applications today. Are there any underserved areas that you feel like AI/machine learning is either not being applied to but would be a great application for? Or maybe areas that these kinds of applications are being attempted but failing. Are there any instances you’ve heard specifically where AI is, for lack of a better term, failing?

M: Well, I wouldn’t say it’s failing. I think it’s just a tougher nut to crack than we initially thought. In the field of marketing, we are trying to pull together customer website and branch behaviors, their product and service utilization, social engagement, and transactional information. Getting the cleaned data into a central data mart is a huge challenge, and it needs to happen relatively fast. I’ve got 3 or 4 different vendors where I store this data and several different formats. It’s tough moving data from A to B to C in a fashion that’s fast enough to be relevant.

For a transaction, I want to do something in real-time. You just made a transaction, and now I want to know which push notification should be sent to you based on an AI/machine learning model’s recommendation. Several third-party providers have the data but are unable to send it in real-time. Or vendor APIs don’t integrate well. Mixing and cleaning data sources together from multiple sources is time-consuming. Poor data quality will make any AI algorithm irrelevant.

I wouldn’t blame AI; we are still fighting many big data challenges.

J: So, would you say that the biggest challenge is data integration? What about data availability or data security? The GDPR (General Data Protection Regulation) in the EU has given consumers of any organization located in or doing business in the EU with consumer data protections and new access controls favoring the consumer over the organization. In terms of data security, usually not a week goes by without another data breach which can significantly impact consumer sentiments. So, thinking about those three areas, is there one that you see causing the biggest headaches in your industry?

M: Historically, banks have had their systems located on-site to manage their own security—there is no cloud. Bank IT wanted everything in on their servers because it makes it easier for them to maintain and protect. That was good, in theory, but very quickly they started allowing more and more cloud service providers. More and more vendors refused to deploy on-prem. Cloud providers have started to address many of the bank’s information security concerns, but banks have very high information security expectations. Banks are based on trust.  But it is hard to deny to flexibility, scalability, performance, and cost benefits of the cloud.

J: Great information, Matt! Thanks for your insight into how financial services are dealing with AI and machine learning.