Wells Fargo, the world’s fourth largest bank, embarked on a massive data transformation journey last year. The goal was to drive efficiency in business processes, improve customer experience and meet the changing regulatory requirements. This comprehensive initiative is phased out over a period of three years. A year into the project, the bank is already reaping benefits from the now centralized data lake.
For a bank the size of Wells Fargo, with over 70 million customers globally, the data transformation project is certainly a highly important initiative.
Prahalad Thota, Head of Enterprise Analytics & Data Science at the US-headquartered bank, shares more insights about Wells Fargo’s analytics and AI/ML programs.
Q: Tell us about the data transformation journey at Wells Fargo.
A: We embarked on a comprehensive digital transformation journey last year, with a three-year plan. Within a year, we have made a lot of progress in terms of the platform, data lake and data environments. We have created a central data lake and about 1/3rd of our data environments are migrated to the lake now.
We have a roadmap for bringing in the rest of the data sources by end of 2021.
While we are at it, we have already started to leverage the data to drive new business use cases, especially around AI and machine learning. Our data policy and data governance structures are in an established phase. This helps my team to ensure that business data owners are aware of what the data is going to be used for. The processes are quite well-established and defined by now, which is a key milestone in this journey.
Q: What are the key business benefits that this project is expected to deliver on completion?
A: There are few broad benefits and the primary goal is to improve efficiency. When data is organized in one place, we will have well-defined processes for access and use of data. For the data analytics team, this would result in improved efficiency while working on any business problems or delivering the use cases.
Secondly, the team can deliver real value for business out of this data --whether it’s to improve customer experience or to manage/mitigate risks.
We are already able to successfully deliver use cases using the existing data from the lake. That said, some of the use cases need to combine data from different sources. Full-fledged benefits will come when we have a completely centralized data environment in place.
Q: Why is centralization of data so critical for the bank?
A: Data environments are typically quite siloed. Data sources are disparate and reside in multiple different data environments. At the bank, we have different platforms and technologies. In many cases, the documentation, which is a critical part of data analytics, was not in shape. When we embarked on the data-driven journey, we had to get all this data in one environment; organize the data in an efficient, standardized way with good documentation and governance policy around that.
Q: Wells Fargo has a comprehensive AI program as well. What are the use cases you are exploring around AI and ML?
A: AI and ML is a very big priority area for Wells Fargo. We have put together a structure to go after the opportunities in this space. This involves a collaborative approach among three core teams of the bank—Innovation, data science and technology. Each of these teams have set out priorities in terms of identifying the right use cases, building the required models and ensuring deployment.
One of the core priorities of our AI program today is creating personalized experience for the bank’s customers. We also focus a lot on technologies and algorithms that help us understand the customers’ interactions and feedback with the bank through various channels. We also have virtual assistants or bots to help users—both customers and employees—in accessing the right data.
Additionally, AI/ML is now being leveraged extensively around risk, compliance, fraud and collection. We are also looking at driving deposit growth for the bank using these technologies. Most importantly, each of these projects have identified particular business problems which need to be solved at the end of the completion.
We delivered around eight solutions this year, based on AI and ML. In 2020, we are going to ramp up our AI program and deliver more than 200 solutions over the 12 to 18 months.
Q: What will be key priorities for you as the head of Analytics and Data Science for the next one year?
A: We have already centralized all our AI and ML model development into one organization within the bank now. I am also looking at standardizing a lot of our model development processes to bring more efficiency. Another key priority will be bringing more automation into what we do.
We are expanding massively in both US and India. As data science skill sets are in high demand across sectors, we are making concerted efforts to hire both experienced candidates and early talent. We are expecting to double our team in next 12 months.