Sponsored By

FICO EVP of Software: Building an AI Platform in Financial Services

Stephanie Covert, executive vice president of software at FICO, joins the AI Business podcast to talk about their AI platform. She also leads FICO's unique women's leadership program.

Deborah Yao

December 20, 2023

15 Min Read
Image of a digital bank
Getty Images

Stephanie Covert, executive vice president of software at FICO, joins the AI Business podcast to discuss the credit-scoring giant's transition to an AI-enabled platform in a highly regulated industry.

She also leads a unique women’s leadership program at the company that has seen 70% of its members see expanded roles, with a third also getting promoted. Covert shares the special approach she uses for this program.

Listen to the podcast below or read the edited transcript.

Tell us about FICO and what you do there.

FICO is a global analytics software company. We were founded back in 1956, so we have been doing this for a little bit of time now, and we are most known for introducing the first commercial credit scoring system. But there has always been two major businesses lines at FICO. The first is our FICO scores business − and most people know our FICO scores business because of the FICO credit score in the U.S. The second business is our software business. The FICO software business is focused on delivering analytics-powered software that helps our customers manage decisions across the entire credit risk lifecycle.

An example of this might be the card originations process. When a consumer goes to a bank, and they apply for a credit card, what the FICO platform does is it helps the bank assess the credit risk of that consumer and approve or deny the application. For me personally here at FICO, I have the great opportunity of leading all aspects of the software business, and that is everything from sales all the way to product development.

I understand that you are leading a transformational shift at FICO to an AI platform. Can you tell me what motivated the company to do this and also what you are doing exactly?

What we have seen over the last many years is there has been a really a big shift in customer expectations. Customers want their organizations to work with them and to deliver value-added and really personalized experiences. (However), there are really a lot of silos that existed in banks, and these silos were driving a disconnected customer experience. They did not really have a true 360 degree view of the customer to be able to deliver those great experiences.

Coming out of the pandemic, the issue has only amplified. During the pandemic, all the interactions with customers went digital, moving from being primarily in the branch to 100% online. This forced the banks to go through digital transformation. And now that they are dealing with their customers in this new digital world, it became even more important for them to be able to deliver a great customer experience digitally, and to make sure that the experience was really personalized, so they could still have that personal connection with their customers.

What we realized was that rather than selling point solutions revolving around solving a singular customer need, what these organizations really needed was a holistic way to view their customers. They needed to be able to look across all of the data that they had about these customers; they needed to be able to use AI to harness the insights out of that data, and then use those insights that they garnered to optimize every interaction with their customer. And because each touch point is important along the journey, no matter what it is, they needed to be able to deliver the right experience every time in order to continue to build that loyalty.

What we have been doing as part of our platform transformation is taking the decade's worth of innovation that we had in our standalone applications, and unify those capabilities into a single SaaS platform. Now our customers are able to use these platform capabilities to build new services and new solutions very quickly to address any needs they have with their customers and each use case that they build they are able to enhance the next use case because all of the use cases share decision assets, they share AI, and they share data. And so each use case that they build on the FICO platform to address the customer lifecycle makes the ones before it even better, and makes the customer experience even stronger.

This transformation is really to help our customers build stronger and more profitable relationships with their customers, but also enable them to really optimize those customer interactions. So they are delivering the right experiences for their customers, and they are building that customer loyalty.

How did you build the AI platform? Did you use open source or partner with a cloud provider and used their tools to build?

We do a combination of the above. Our cloud platform is built on the infrastructure of Amazon Web Services today. The capabilities are a combination of leveraging some capabilities from open source, but a majority of the capabilities that we deliver are capabilities that have been built by the FICO team. They are very unique; they are very differentiated. In fact, we have a lot of patents around the capabilities that we built. So it is really a combination of an open source of FICO-specific IP, and then leveraging a public cloud infrastructure to deliver that to our customers.

… Our platform is used by over 100 of the top Tier 1 banks in the world today.

Some of the most popular uses on our platform are to drive analytics-enhanced decisioning. Our customers come to us to solve a number of different problems, whether it be an originations problem − they might want to make better, smarter and faster decisions on the originations front − or they might want to be tackling better customer management and they want to be able to use all the data that they have about a customer to be able to make smarter decisions, whether that is how to market to them, when to make an offer of maybe a credit line increase, or how to treat them if they fall into hard times and to go into pre-delinquency, or collections.

There are hundreds of use cases for our customers on the platform, but they are really rooted in bringing in data into the platform, applying analytics to that data, finding the value in that data, and using those insights to drive optimized decisions. And then using the feedback loop and the learning mechanisms in the FICO platform to make each and every next iteration smarter.

We do not use generative AI today on our platform, but we have the ability to enable our customers to use generative AI. What we do is make sure that we can help them use that generative AI in an auditable and explainable way. And so it is really about them being able to take a model that they might have built in-house and execute that as part of maybe a marketing strategy or another strategy within their portfolio. But we make sure that we can understand how those decisions are being made and they can always point to, and report back on, the audit trail of what was being done around that decision.

Is there a reason why you are not offering generative AI at present?

We do think there is enormous value for generative AI inside of financial services. But it is a highly regulated environment. What we want to do is make sure that any AI that is used especially in the financial services sector is responsible. Our focus right now is really on the R&D around how do we make the generative AI responsible, auditable and ensure that it is used in an ethical way, when it is used on the FICO platform. That is where we put our energy in terms of looking at how we would bring it to our platform.

Since you deal with banking clients, what is your sense about their appetite for generative AI? Is everyone a little bit cautious about it, or they are very gung ho?

I think you get the entire spectrum on that: Some people are really excited about figuring out how they can use some use cases right now and some are a little bit more cautious. Some examples of where our customers are interested in exploring generative AI is really around how can they improve their consumer interaction. That might be in the areas of customer service where they want to enable smarter chatbots to help customers self-serve, or in the area of communicating with customers, they want to better understand … not only their preference of communication channel − do they want me to give them a call, send them an email or send them a text − but when is actually the best time to reach out to them that they might be able to interact and resolve an open issue or respond to an offer that I am sending them.

So they are really looking in the areas today around that customer experience and customer support perspective. But I also think everybody in the banking industry does have a little bit of caution because they do they do want to make sure that they are using it in a responsible way. But I know that people are excited to start to find really valuable uses in the market.

How is FICO using AI internally?

From a FICO perspective, specifically we embed AI directly into our solutions. Customers can leverage FICO-built models in AI right in the FICO platform. But we have also built a number of AI development and execution tools as part of our FICO platform. We are focused on ensuring that our customers are able to operationalize their AI (projects), bring them into production and ensure that they are getting value from their own investments. We are very happy for customers to leverage our models and our AI in their decision-making processes.

I would also say that we spend a fair amount of time on research and development when it comes to AI. Today we hold around 218 patents, all pertaining to the area of AI. This spans areas from interpretable machine learning to ethical and auditable AI. We are using AI in every solution that we build, but we are also enabling our customers to bring their own as well.

What are some of the biggest challenges you see your banking clients encounter when they are trying to deploy AI in their own companies? And also, what are some of the solutions that you have seen?

Our customers are really challenged with being able to take AI from the development stage to being able to use it into production. If you look at the statistics, I believe Gartner says about 90% of AI projects fail to deliver the value that they were intended to deliver because of that difficulty in the actual operationalization of it. We see our customers challenged with that every day − challenged with getting the right data in and then challenged with then operationalizing the analytics itself.

We help our customers cut through the problem of, first, getting the data. They can bring in data from anywhere: third party data, first party data, streaming, at rest, structured, unstructured, it does not matter, they can bring that all into the FICO platform. Then with our execution capability, they can use their AI models and execute them in production. We can help them monitor, explain and report on what is going on with those analytics so that they can ensure that they are using these analytics (optimally) and in an ethical way.

Taking about data, fintech startups have been trying to crack the financial services market for years. They use some very interesting metrics to, for example, assess the credit worthiness of consumers, including social media data and other non-traditional information. What does FICO think about this alternative data?

A lot of fintechs talk about using social media data, but the fact remains that the most predictive data for credit performance is really past credit performance. When I think about a fintech maybe using a Facebook profile to make a decision about credit worthiness, we can all agree that might get us into a little bit of trouble. But what I think we are really talking about is leveraging alternative data and leveraging consumer permission data to enhance credit-decisioning.

In our FICO scores business, we have innovated on scores that use not just the credit bureau data, but also leverage customer permission data (data consented to be shared with a party), or open banking data to enhance the risk-decisioning process. We are fully supportive of leveraging as much data as possible in order to make the best decision possible for customers. But our aim is to ensure that the data being used is producing fair, unbiased and ethical results. We need to make sure that we are not just using one piece of data to determine a credit risk profile.

Switching gears a bit, I understand that you are very passionate about your women in leadership program. Tell me more about that.

It is a program focused on teaching women to lead their own self first, so that they can lead others. We focused on developing the skills that are needed to be an authentic leader, and to really know yourself, so that when you lead, the base of values that you are leading from comes across in an authentic way.

We have had a tremendous amount of success in the program; we have had over 70% of the participants seeing their responsibilities expand since completing the program. And over 30% of the women at this point have received a promotion after going through the program. What I am most proud of is that the program has created a community and connected the women inside of FICO because women in tech is still a small group. We have really been able to create a large global community of women that are supporting each other, encouraging each other, and mentoring each other so that they can succeed together.

Why do you think there continues to be a lack of gender diversity in tech?

You know, that is a really great question. Women are entering the workforce at a higher rate than men today … and they comprise 50% of the workforce but only 26% in the tech industry. Why lack of diversity still exists, I would say it starts with there being a lack of diversity in leadership roles. That means there is a lack of role models for women. The makeup of the leadership team really matters because it signals to individuals that are looking to join an organization or even looking to reach for higher positions in an organization, whether they can reach those leadership ranks and whether they can succeed. And so, the leadership itself is kind of discouraging to some people.

Second, from a technology perspective, in general, we are not encouraging enough women to enter into the areas of STEM. And among those women who do, half of them still do not join the tech industry. A lot of them go into research or other areas. We have a lot more work to do up front to make sure that we are encouraging women to come into the space and to the industry, and that we are making sure when they do that, they are set up for success.

What can be done to have more women in top leadership in tech, just like yourself?

First, we need to have more mentorship roles for women. Most women in the workforce, if you were to survey them, would say they have never had a mentor or a role model in their career that was helping them to succeed. Men are mentored at a higher rate than women and this puts women at a huge disadvantage. So when an opportunity arises, sometimes women are not prepared or they have not been given the skills necessary to have the skillset for that job. If we can improve mentorship, that would (help women rise to the top.)

Second, we still need to think about the environment that women are working in. Women still face challenges. Women are still suffering from microaggressions; we are still suffering from harassment in some workplaces. And there is a lot of gender bias still out there. Even men are often seen as more competent and more capable than women in leadership and women are often overlooked.

If we want more women to succeed, we need to create an environment that is psychologically safe for them, and it is supportive and unbiased, where everyone can thrive and survive and everyone has an equal opportunity to succeed.

The last thing, and this is really important and it is why I started the FICO women leaders program, was that we really need to do a better job of teaching women about leadership skills and about self-awareness. Women and men, we tend to at times have different leadership styles and qualities. (But) far too many women are really pushed to adopt what would be considered maybe a more masculine style of leadership and told that they need to be a certain way in order to reach higher levels in an organization. They are really not taught to embrace their strengths.

Women leaders are often more empathetic; they often make decisions from a sense of community and not a sense of self. It is that diversity of thought, and that diversity of approach, that makes organizations stronger. If we can teach women to be themselves and be the best version of themselves, and to operate from a place of values, we will be able to create the self-confidence in women to want to reach for those positions and to want to lead others.

Read more about:

Applied Intelligence

About the Author(s)

Deborah Yao

Editor

Deborah Yao runs the day-to-day operations of AI Business. She is a Stanford grad who has worked at Amazon, Wharton School and Associated Press.

Keep up with the ever-evolving AI landscape
Unlock exclusive AI content by subscribing to our newsletter!!

You May Also Like