Dr. Jose Murillo, Chief Analytics Officer at Grupo Financiero Banorte, speaks with us about using AI for revenue generating projects as well as cost cutting strategies to further digital evolution.

Avenue Code: Tell us about your personal career path. How did you become the Chief Analytics Officer at Grupo Financiero Banorte?

Dr. Jose Murillo: After graduating from Rice University with a Ph.D. in Economics, I joined the Bank of Mexico’s research department, working on several strategies to mitigate the impact of the financial crisis. I eventually became the Director of Information, Price Analysis, and Regional Economies and held the staff’s view on the main decision variables for the Monetary Policy Committee.

When the governor of Mexico’s Central Bank started working for Banorte, he invited me to come as his advisor. A few months later, Mr. Rafael Arana, a newly appointed Chief Operations Officer, shared with me his vision for an analytics business unit and asked me to build it. He wanted to establish the unit as a profit center, which is a very unusual focus for data science and analytics teams in North America. 

In my first year, I produced 46 times the cost of the department, and last year, with more data scientists on my team, we made 245 times our cost, almost 1 billion dollars in net revenue. During my tenure at Banorte, we’ve seen more than 3 billion dollars in net revenue for the company, which has allowed us to become the most profitable bank in Mexico, the second largest financial group in Mexico, and one of the most profitable banks in North America.

AC: How has data enabled you to cut costs so dramatically? 

JM: On our transformation journey to become a data-enhanced financial group, the emphasis was initially on cost cutting strategies for operational and financial risks. This was the focus of almost 50% of our projects. Progressively, revenue generating projects became much more important. We’re now primarily focused on increasing the value of our business relationships with customers. Data is the basis for our strategy to enhance customer experience because it helps us deliver products at the right price, at the right time, through the right channels.

Beyond this, AI helps us accelerate the learning curve. When we test new interventions, we start small, then scale up. AI techniques help us generate results faster. 

AC: What are the biggest challenges and opportunities you have faced as CAO of Banorte?

JM: The biggest challenge is overcoming skepticism. Many companies that want to implement data science, analytics, and AI begin by partnering with external companies. What sometimes happens is that there is not a close enough co-creation process between the business and the data scientists, which means that the technology is not used to solve the right questions, resulting in a low ROI. But as soon as you’re closely aligned, you can prove the value for the technology and foster a willingness to transform.

When it comes to opportunities, it’s important to look at revenue generating projects as well as cost cutting initiatives. You have to have some degree of credibility for revenue generating projects. For example, perhaps you know that you’ve sold X credit cards, but you don’t know the long-term impact - will those credit cards have a good or bad record? Are they going to perform, or not? You have to agree on metrics for success, and those metrics must correlate to real business value.

At the end of the day, revenue generating projects must help you extend the duration, depth, and width of the business relationship with the customer, which is what allows you to increase customer lifetime value. The value of a company is based on its business relationship with its customers. 

AC: How do you overcome skepticism and show the potential for data-enhanced transformation?

JM: Implementing data science, analytics, and AI mandates transformation at the basic level, which means breaking silos and restructuring. It’s not enough to have strong analytics or strong data science expertise. You need to be able to effectively communicate your goals for these technologies, as well as the expected results - that’s how you make a case for their value. You also have to create good will between teams, giving credit where credit’s due. 

Finally, you have to establish the right incentives and ensure alignment between individual teams and the corporation as a whole. At Banorte, we present projects and vote on which ones have the biggest ROI potential. This way, everyone sees that if the projects are successful, they will add value to the company, which benefits everyone. 

AC: What is the future of AI in the financial sector?

JM: AI is one of the most fascinating developments in finance. While AI has been present in academia for several years, firms are only beginning to understand its potential. In the financial sector, the goal for AI is to help build a long-term relationship with customers. For example, the goal for both firms and customers is for the latter to gain financial stability over time. That means they need to start saving now, but there are several human biases, like hyperbolic discounting, that act as barriers. So firms can help customers, and there are a lot of results that are well-established. But the problem is determining which of these are relevant to the business and how to communicate them efficiently to the customer. The way that you understand that is through small, well-designed interventions.

In this particular situation, AI helps us understand how each small intervention works and predict which interventions are effective to scale for a large population, accelerating the learning curve and enabling companies to meet financial targets. In other words, AI assists corporations in better understanding the customer. 

AC: What are the benefits, challenges, and limits of AI in the financial sector?

JM: AI helps us better understand the customer, understand the risk, and improve risk models. This, in turn, helps us to significantly expand our customer base.

One of the challenges for AI is the black box problem. We can change this by explaining AI techniques and showing their value. On the marketing side, AI can also help build a more efficient process for customer relationship management. Once you can understand the customers better, you can define better products and deliver these products at the right time, at the right price, through the right channel. These are the main advantages.

One of the reasons it’s hard to prove the value of AI is that it’s difficult to demonstrate an impact on ROI. Part of the problem is that people are too focused on using AI for cost-cutting strategies like chatbots. On the other hand, I think it can be used to enhance and humanize the relationship with the customer. With AI, you can be much more efficient in having meaningful conversations with customers and much more attuned to what the customer is saying. You also avoid wasting the customer’s time. Using AI only to save money is problematic. 

The limits of AI are primarily ethical. This technology should be used with a sense of responsibility and with a receptivity to feedback and criticism. 

AC: What are the current data trends, and how do you plan to adopt them at Banorte?

JM: Value has to be demonstrated. If data science and analytics groups cannot prove value, AI won’t be used. The biggest trend right now is to be able to measure the impact of your work. This is one of the most important things for the industry, because data science investments are expensive, so if value is not proven, it will be a passing fad. 

On the other hand, the industries that understand its potential will move at a quicker pace. In that sense, there’s going to be a bigger focus on how data science helps boost revenue instead of just cutting costs.

AC: Thank you for the insight into how AI, data, and analytics are being used in the financial sector, Jose!