200,000 reasons to thoughtfully integrate AI: Q&A with Wells Fargo’s AI chief

At a time when most laypeople were chuckling in social media forums over bizarre AI-generated images, Andre Mansour saw opportunities.

Processing Content

A self-professed “big planner” who enjoys doing research, Mansour has a longstanding comfort level where AI deployment is concerned. Over time, it has become part of both his personal and professional life and a useful tool that makes both better.

Mansour recently joined Wells Fargo’s Wealth and Investment Management (WIM) team on the heels of nearly a decade at Google, where he oversaw investment banking at Google Cloud. His responsibilities at Google included the leadership and delivery of solutions that reduced expense ratios at large, global financial firms.

As they embrace AI integration, firms of all sizes have begun to understand the importance of finding experienced leaders in this relatively new, ever-changing industry. Sleuthing out veterans who can best interpret the tools to make things run efficiently is part of the criteria. Avoiding past mistakes that have cost other firms is another.

In addition to his experience at Google, Mansour brings both enthusiasm and calm to Wells Fargo and its 200,000 employees the AI integration will impact, even as WIM manages $2.2 trillion in client assets. From a professional standpoint, he emphasizes the company’s “top-down commitment to allowing AI to deliver a new way to approach banking, wealth and investment management in a governed way,” he said. “It’s a core competency that we are building and will deliver here.”

His own relationship to AI is one in which humans and machines also integrate to make his personal life run more smoothly and efficiently.

“Most importantly,” he said, “it allows me to spend more time with my family, feeling like I have everything I need to get done because I have assistance in a new way.”

Mansour spoke with Financial Planning about his new role.

This conversation has been lightly edited for length and clarity.

Financial Planning: Congratulations on your new role. Tell us what you’re excited about as you settle in and get started.

Andre Mansour: I spent about a decade or more in my career as a practitioner within the space. It really gets me excited to bring Silicon Valley insights and mix that with the domain knowledge I learned earlier in my career and bring that together to deliver something that’s going to be useful for Wells Fargo and the customers.

We have a diverse set of business practices that provide a full portfolio for our end customers, and a deep reservoir of data and information that will enable us to excel in that. That’s why I came here. That’s what makes me the most excited.

FP: What is going on in AI at Wells Fargo?

AM: Our philosophy here is around using AI for outcome focus. We’ve invested an enormous amount, from the top down, in deploying and modernizing our technical infrastructure in a meaningful way that is focused on AI-related outcomes. We’re deploying AI across our branches and customer service centers in advisor interactions, and to remove any friction and fragmented systems.

READ MORE: How RIAs can prevent serious AI fails, reduce regulatory risk

FP: Let’s get the pain out of the way upfront. What is a challenging part of the job?

AM: The hard part is really allowing for the financial advisor to have the right information at the right time to do their job. That’s hard.

FP: This also presents an opportunity.

AM: There is an unlimited number of ideas for AI and its nomenclature for everyone at every level of technological sophistication. And our opportunity is to identify the 20% of initiatives that will drive the 80% of the opportunity for us to be productive for our investors, customers and the platform’s internal users. So it’s really about having an enormous amount of discipline for the use cases we want to put in place. And then to have the technical capability in-house to deliver those.

FP: Give us some examples of how AI is currently used.

AM: It is being used in the Intelligent Banker book, which provides a unified AI overview of customers from a CRM perspective. Another one is a tool we built around the credit memo desk, which uses AI to reduce the time required to prepare credit memos, drive consistency, reduce risk and improve delivery turnaround times.

Finally, we have a platform called Advisor Gateway. It’s the control center for all advisors in our practice. It has over 200 tools and applications. We are delivering AI within that platform in various ways that we’ll announce more in the future. We have a phased approach, but it’s moving rapidly. 

FP: As you wrap up implementing components like notes drafting and copilots, what else is next?

AM: We are moving toward having insights into when it’s important to have conversations with the client. Do we understand in context the balances, products, etc., that are appropriate for the client’s situation? Next, we’ll deliver agentic systems, providing long-running assistance for things like changing beneficiaries or money movement.

READ MORE: Bank of America gives Merrill Lynch an AI makeover

FP: We know AI moves fast, but where do you see things headed over the next three to five years?

AM: There are a number of different layers of it. There’s the front, middle and back office. When you move from this horizontal set of productivity capabilities that can traverse from front to back office to things that are more about contextual delivery, the opportunity becomes about surfacing the right data at the right time for the right user.

And that is an opportunity for every major enterprise attempting to deliver this because the models are, effectively, commoditized, or there’s small, incremental differences amongst them. But what really matters is how they are presented and traverse the corpus of information within your organization. And thankfully, Wells Fargo has a real and unique differentiation in that area, having such a wide breadth of businesses and depth of understanding and partnerships with their clients for a long period of time.

We are ending the productivity phase and moving to where the data and workflow are in place. The coming three-to-five years are about long-running agents. So, back to that analogy about the front, middle and back office. For a particular workflow to traverse the front, middle and back office, Wells Fargo is working to have a cleaner, fresher standard of data, a very clear and well-informed workflow, and the ability to deliver an agent that can run over a long workflow.

Think about how many steps it might take to change a beneficiary or even simply change a password. There’s a number of steps there. To have an agent traverse that, we need the set of risk and compliance guardrails in place, and an agent smart enough, with the right information and data in hand, to complete that set of tasks, always with a human in the loop at the end or in the middle. But they’re able to make multiple decisions over time on behalf of the user.

FP: In terms of feedback, what are advisors telling you? What do they like, hate and ignore where AI is concerned?

AM: We bring our advisors into the design process early and often. It’s a must-do for us to be successful, and it’s about adoption metrics and usefulness.

So what do they like? They like when things are easy and useful, and delivered in a conversational way, which is what generative AI can provide. It’s really good with unstructured information and, as a conversationalist, surfaces insights in any question you ask it, within reason.

What do they hate? When it’s outside of their workflow, which is exactly why we delivered a platform like Advisor Gateway to have everything in one place. To deliver a silo tool means they need to jump to another application in another place, which is not useful and something they don’t use.

What do they ignore? They ignore it when it’s not about the end customer: their clients. If it doesn’t ultimately allow them to focus more time on their clients and their clients’ financial freedom and positioning, it’s not important to them.

FP: We often hear about AI replacing humans. How are you handling that?

AM: The core of all of this is that humans will always oversee the process. It’s not about replacing human judgment. It’s about helping advisors and associates spend more time helping clients achieve their end goals and not doing the busywork. A key design principle is to have a governance system with human oversight and embedded risk and compliance guardrails.