For Wells Fargo, there isn't a defined approach yet to integrate evolving technologies such as artificial intelligence into banking.

Each idea comes with different potential benefits and costs, and banks need teams that can work in an environment that is safe and separate from the standard thinking of the main organization, said Shari Van Cleave, head of Wells Fargo Digital Labs.

With a team of 30 people who are business strategists, technologists and designers, Van Cleave focuses on driving innovation in Wells Fargo’s digital banking.

Previously, she was a founding member of Prudential Financial’s strategic venture capital and innovation lab. She also held senior-level marketing and product positions at World Vision, a global humanitarian organization.

Van Cleave took up her current role with Wells Fargo last May. She sat down with American Banker recently to discuss what the digital lab does, the ups and downs of integrating artificial intelligence, its new pilot of Apple Business Chat and plans underway for the use of augmented reality.

The interview has been edited for length and clarity:

On an average day what do you spend most of your time doing?

SHARI VAN CLEAVE: I don't know if there is an average day. One day we could have customers that are in our lab playing with some things that we've recently developed to get their insights and determine how they would use it in their lives. Another day might be meeting with fintechs.

One of the great things about being in San Francisco is we're right there in the heart of the tech community. We can determine if this is the right time for us to bring those to our customers or experiment with those partners.

It could also be meeting with big tech firms. Recently, for example, we worked with Apple to launch Apple Business Chat. We were one of the first major banks to do so in March.

Another day might be looking at our own business model and how we're operating today, and asking tough questions about ourselves.

How will Wells Fargo use Apple Business Chat?

This would look a little bit different than a normal conversation. You have Wells Fargo's logo on top so you know it's us. You can text and ask us anything. Such as, is there a branch near me? This question will go to a banker on the other side of this conversation.

I don't think any of us anticipated the degree to which texting would change our lives. We now text each other more than we call each other. But we haven't done that with corporations. We continue to call them. Who actually wants to call a corporation for customer service?

There's another component that really got us interested in this particular channel, which is that we can embed our own technology in the interaction.

If I were to go into a more detailed conversation with the banker, they would need to authenticate me. They would need to be able to say if this is Shari. We've been able to build this technology into our own app experience, without the user needing to leave the conversation. It authenticates that this is Shari so that the banker knows who I am.

You can imagine in the future that you're applying for a home mortgage, and you sign a document in a text conversation with the person who is the provider of your loan.

"Who actually wants to call a corporation for customer service?" asks Shari Van Cleave, head of Wells Fargo’s Digital Labs.

Will AI eventually be integrated into the service?

We imagine that there might be bots or other things at the other end in the future that are integrated with the human service at the right touch point.

We want to be balanced with creating a human-designed experience, but we also realize customers want something really quick and simple. So that can maybe be a bot that can answer one question and a human that answers another. Right now, as we start this, it's just a human.

We're really excited about when this goes national. Right now, we're doing a pilot in three states. We want to learn how customers use it before we roll it out broadly to the whole universe. Right now, it's in a test space. Later on this year or early next year, we imagine it going national.

What results are you looking for before going national?

Well, the first thing is to answer what are the use cases that people want to have answered here. What are the kinds of questions that they might come here with, rather than to the mobile app or a phone call. Then that enables us to build technology around those kinds of experiences.

One of the first uses cases that we feel is useful is when you've lost a debit card. Let's say that I got here this morning, went to get coffee and left my debit card somewhere in San Francisco in a rush. I could just text someone and say, "Hey, I've lost my debit card. Can you help?" And they could just turn it off. That's an example of a use case and the technology behind it. We're interested in what other use cases we need to design around before we go national.

In the past, you’ve mentioned that at Wells there's technology that monitors bias in AI created by the data that humans are feeding it. How does that technology work?

Well, let me preface this by saying it is about the people, first. It is about setting the right goals of what we have, so we know that AI is measuring what we want it to deliver. We have to make sure that those goals are the right goals.

Are we setting diverse targets for our communities? Also, are we setting diverse leaders within that space?

Another way is by finding a way to remove the bias from your training data for a model. I think another example on the back end is having models that check the models that you have. Models, for instance, that interact with one another.

Some of the things that we see with models are that they are really specialized. Some models are set out to answer a specific question, but they sometimes go out to answer that question without the completeness that you want. So if we have another model that says, "Wait, let's come in, let's check that, or let’s ensure that your data is complete," you can then go back and run a second model that’s more individualized.

How does what you do relate to Wells Fargo’s Innovation Group?

At large organizations there are a couple different ways you can organize around innovation and some entities take it at an enterprise level only. Some entities will say, let's have different innovation groups aligned for each of the business lines. At Wells we say it's both/and. It’s an ecosystem.

So the Innovation Group works across the entire enterprise and looks at really cool, far-out questions. They deliver capabilities across the entire enterprise and we partner with them to bring those technologies and capabilities to our areas of practice. I'm the lab for our digital group. So I go across the enterprise to ask how digital banking could change our business.

What do you plan to do with augmented reality?

I think playing with augmented reality is another way of looking at data visualization. This is why we have designers on our team and why design works really closely with technology.

Some things that we've taken for granted is that we see data in a table. Design is allowing us to create more human and empathetic relationships with our data. Whether it's color that make me feel more soothed. Or how do I make looking at your balance a not scary experience? Can we interact with numbers in a different way for people that see or understand math differently? Imagine a space where you're walking around seeing it and that might help you process it differently.

Something like this is where we want to ask questions that are provocative to see what the future could be maybe three to five years out.

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.