Conferences, AI Snake Oil and the Future of Fintech

Hey everybody. Thanks for joining us for another episode of What’s Going On in Banking. I’m Ron Shevlin, chief research officer at Cornerstone Advisors and author of the FinTech Snark Tank on Forbes. And, of course, I’m here with my co-host Stacey Bryant.

Stacey, how are you doing?

How are you doing, Ron? I’m fan-flipping-tastic.

I’ve been on the road this week, and it is definitely strategic-planning season. I’m getting calls, LinkedIn DMs, texts, you name it. Everybody is trying to understand who is doing what and how they should respond.

I know you’ve been on the road too. FinovateFall happened earlier this week in New York City, and there’s a lot to talk about across technology, fintech, and AI.

This may be the “on the road” edition of What’s Going On in Banking, and it could stay that way for the next couple of months.

I was at the Jack Henry Connect conference in San Diego. I ran into a lot of people, and every conversation started with, “How’s it going?” My answer was basically, “Summer is over and we’re back on the road.”

The frustrating part is that so many good conferences overlap with each other. I had already committed to a couple of things at Jack Henry Connect, so I couldn’t make Finovate.

You were there for at least part of it, right?

Yeah. I was there Monday, which was the first full day.

I’m a New Yorker, even though I live in New Jersey now, so anytime Finovate is in town I try to show some love. Shout-out to Katie and Greg Palmer for having me and our colleague Sam Kilmer there.

The event kicked off with something that really hit close to home for me: a women-in-financial-services gathering.

The room had baby boomers, Gen Xers, millennials, and even Gen Z. There was a 19-year-old woman there who wanted to understand how to get support for her startup, whether that meant funding or just mentorship.

That kind of representation matters.

I also ran into Sherry Wu from the University of Michigan Credit Union. She’s their chief technology officer and has been there for roughly eight years. She came into the organization without a traditional credit-union background, and her approach has been to ask questions, challenge assumptions, and bring diversity of thought.

She made a good point about vendor pitches. Right now everything is “AI-powered this” and “AI-powered that,” so her question is simple: what problem are you solving?

Those are the questions you and I have been talking about for a while.

Then I sat through a lot of demos.

I’ve been a presenter at Finovate before, so I know how difficult it is. You have to be on. Some demos went south. Some were just okay.

The ones that didn’t work for me usually failed for the same reason: I couldn’t immediately tell what problem they were solving. I could see an interface, but I couldn’t identify the pain point quickly enough.

One demo that stood out was Swaystack.

Har Rai Khalsa, who’s been around this space for a while, did a very disciplined demo. He took his time, used nearly the full seven minutes, and clearly explained that Swaystack is focused on reducing attrition and re-engaging accounts that are going dormant.

As he explained the problem, integration, and implementation timeline, I could see more people in the audience taking notes.

That told me something important: if you clearly define the pain point, show how the technology integrates, and explain how long implementation takes, the audience pays attention.

Debbie was another company that caught my eye. It focuses on financial wellness and financial literacy. I’m seeing more and more of those kinds of tools.

And, of course, Sam Kilmer spoke about how institutions are looking at their tech stacks, how they can begin experimenting with AI, and how data strategy fits into the whole picture.

That was my Monday at Finovate.

I wasn’t there, but surprise, surprise, I have a couple of opinions.

First, I like that you called out Swaystack, not just because of the company, but because of the bigger trend it represents.

Traditionally, bankers have thought about the customer lifecycle as acquire, cross-sell, and service. Sometimes the order changes, but it’s basically those three things.

What Swaystack represents is recognition that there is another major stage: onboarding.

You can acquire a customer, but then you need a period of relationship-building.

And onboarding is not just the first seven or 14 days after an account is opened. It can be a multi-month process, maybe even a year, where the institution makes sure the new customer is actually using the relationship.

That means direct deposit, online banking, mobile banking, bill pay, money-management tools, and other features.

So I’m glad to see onboarding becoming a more explicit part of the relationship-management model.

The other thing I always watch at Finovate is the Best of Show winners.

My first Finovate was in San Francisco in 2007, so I’ve been watching these awards for a long time.

And for years, I was pretty critical of the voting because bankers seemed overly impressed by shiny front-end user interfaces.

I still remember one year when MX demonstrated bubble charts and won Best of Show. I went ballistic on LinkedIn.

My friends at MX probably won’t love me bringing that up, but my point was: who cares about bubble charts?

Bankers were voting for interface features instead of the technologies that could really transform banking.

This year’s winners were much more interesting to me because of the problems they addressed.

One was Casap, which focuses on card and payment dispute resolution, particularly around fraud.

I was pleased to see that kind of company win because it reflects growing recognition that dispute and fraud resolution are strategically important.

By the time people listen to this episode, Cornerstone may have already published research we did on the fraud-dispute experience. The study was commissioned by Quavo, a Casap competitor, and surveyed 2,500 consumers who had gone through the fraud-dispute process.

The findings are important.

Consumers generally don’t blame their financial institution for the fraud itself. They understand that someone else committed the fraud.

But they absolutely judge the bank or credit union on how it handles the dispute.

The speed of the resolution, the quality of communication, and how well the institution brings together the right information all affect satisfaction and the customer’s willingness to deepen the relationship.

With fraud levels rising, more customers are going to have these experiences. That makes fraud resolution a relationship differentiator.

Another Best of Show winner was Eko, which is integrating investment capabilities into deposit and checking accounts.

Again, I don’t want to sound like I’m just promoting Cornerstone research, but we published a report recently that made a very similar point.

For younger consumers, integrating investment capabilities, including traditional securities and crypto, into a deposit account can be a meaningful differentiator.

That’s especially relevant for community banks and credit unions.

We also estimated the potential impact of what I call deposit displacement.

I started using that term about 10 years ago when fintechs began taking more of the customer’s financial relationship outside the bank.

When interest rates fell, people stopped paying as much attention to it. Then rates rose again, and the issue became much more visible.

Over the last five years, we estimate that more than $3 trillion has moved from banks and credit unions into fintechs and other non-bank providers.

Fighting that outflow is going to require more than turning up the analytics dial.

It requires new products and new capabilities that give customers a reason to keep more of their financial lives with the institution.

That’s my Finovate rant.

And for Eko, what we’re really talking about is investments-as-a-service for a younger demographic, right?

Exactly. Similar to what Investi is doing, the goal is to integrate investing directly into the deposit relationship so a community institution can offer something that stands out.

Look at Robinhood. I recently said they had $18 billion in deposits and someone corrected me because, apparently, they gained roughly $18 billion in deposits just in the first quarter.

That’s a huge amount of money leaving the traditional banking system.

Community institutions can’t simply plead with customers to keep their money there. They have to fight back with better products.

And this connects directly to strategic planning.

Institutions are asking how to grow accounts, improve engagement, and generate more fee income. The Finovate winners give them examples of capabilities that could help move those goals forward.

So virtual applause for the winners.

There are two other things I want to mention from Finovate and Jack Henry Connect.

When you mentioned the women-in-financial-services event, I thought you were going to bring up the tribute Finovate did for Barb MacLean.

Yes. Absolutely.

Barb passed away a few months ago. She was a huge presence in the fintech and banking community and was known for her FinTech Playlist, where she curated industry stories alongside music every week.

I knew her long before that, and she was always a huge supporter of my work.

Her death was a major loss to the community, and hats off to Finovate for recognizing her.

The other story happened at Jack Henry Connect.

Wade Arnold from Moov and Ben Metz were announcing a partnership around a new local-payments product.

During the conference, there was a school shooting in Colorado. That school is where Wade’s children attend.

He obviously needed to get home immediately.

United Airlines couldn’t get him out until much later, so the Jack Henry team stepped in. They got their corporate pilots and corporate jet involved and helped Wade get home hours earlier than he otherwise could have.

I want to give huge respect to Greg Adelson and the Jack Henry team for doing that.

And this is what really frustrates me when politicians constantly bash banks and financial-services companies as if the industry is inherently evil.

This is a community. These are people. And sometimes that side of the industry gets completely ignored.

Dr. Martin Luther King Jr. said that adversity reveals character, and I’m paraphrasing there, but that’s what comes to mind.

At the end of the day, financial services is still made up of human beings.

I also had a personal memory of Barb.

About a year ago at Finovate, I brought a box of Levain cookies. If you’re a New Yorker, you know these things are enormous. They’re ridiculous and delicious.

That was the first time I met Barb in person. She was wearing this great motorcycle jacket, and we ended up talking while she ate one of the cookies.

She told me about the politics of trying to push innovation inside a progressive financial institution. Her attitude was basically, “If I see the opportunity, I’m going to go after it. I can get slapped later.”

That stuck with me.

Rest in peace to Barb.

All right, shifting gears.

I came across an article a few weeks ago with a very catchy headline: “99% of AI Startups Will Be Dead by 2026, and Here’s Why.”

The article argued that most AI startups are going to fail because they’re fragile wrappers around rented models and are dependent on infrastructure from companies such as Nvidia and OpenAI.

It compared the current environment to the dot-com boom and described a lot of startups as companies that are simply putting a new interface over GPT or another model.

That got me thinking about the number of vendors today who say they are “AI-powered.”

I even had ChatGPT create a parody image of Smokey Bear saying, “Only you can prevent vendors from selling you AI snake oil.”

So here’s my question: is there such a thing as a sustainable AI business when the underlying technology changes this quickly?

I have a few reactions.

First, I’ve been a technology analyst for a long time, and one thing I’ve learned is that you can be right 90% of the time by predicting that a startup will fail.

That’s easy.

The hard part is identifying the winners.

Second, the label “AI startup” is becoming increasingly meaningless.

If nearly every technology company eventually has AI embedded in the product, then saying a company is “AI-enabled” doesn’t tell you much.

It reminds me of the early PC era.

At one point, saying something was a “PC application” distinguished it from mainframe or minicomputer software because that was where the new investment and innovation was happening.

Eventually, the label became pointless because PCs were simply part of the environment.

I think “AI-powered” is heading in the same direction.

I may lose a few friends for saying this, but I recently saw a venture-capital firm publish an investment thesis with “AI-enabled companies” at the top of its priority pyramid.

My reaction was: everything is going to be AI-enabled.

That doesn’t tell me where the real strategic value is.

The more useful question is where in the technology stack value will accumulate.

There is AI infrastructure: models, orchestration, agents, security, privacy, and all the things users may never directly see.

Then there are applications built on top of that infrastructure that solve actual business problems.

The article’s prediction that 99% will fail is almost a throwaway line. The interesting question is which layers and use cases will create sustainable value.

And there’s a strategy lesson here that goes back 40 years.

What was Microsoft’s first great product? MS-DOS, the operating system.

But Microsoft understood that dominating the operating system alone was not enough.

It built applications such as Word, Excel, and PowerPoint on top of the operating system. Those applications made the operating system more valuable and helped lock users into the Microsoft ecosystem.

Then the company expanded further into higher-level solutions.

The lesson is that competition and profits tend to move up the value stack.

You can make a lot of money being the infrastructure provider for a period of time, but infrastructure eventually becomes more standardized, prices decline, and competition pushes upward into applications and solutions.

That’s important for banks too.

Banks shouldn’t be obsessing over the lowest layers of AI infrastructure. Someone else will handle most of the models, privacy, security, and orchestration.

In banking, I think of the stack this way: the bottom layer is infrastructure and security. The middle is money movement. The top is optimization.

That’s why money management is ultimately more important than money movement.

Stablecoins will have infrastructure providers. Other companies will build the rails that move the money.

Banks create their most durable value by helping customers optimize what they do with the money.

That’s the higher-value layer.

My last thought is that I’m hearing a similar question from community-bank and credit-union executives all the time: should we build or buy?

One credit union recently considered spending thousands of dollars on a company to build a website. Meanwhile, we know that tools such as ChatGPT can now generate functional website code very quickly.

But institutions are also worried about liability, support, security, and what happens when something breaks.

So the technology changing doesn’t eliminate the need for judgment.

The same thing applies to banking. The important question is not whether AI can technically build something. It’s what the market needs, what behavior the customer is showing, and whether the institution can support the solution responsibly.

We’re going to keep talking about this, and I’m sure we’ll revisit it three more times.

Stacey, thanks a lot. And thanks to everybody listening.

Please look for the next episode of What’s Going On in Banking.

If you enjoyed today’s episode, hit the follow button on Spotify, Apple Podcasts, YouTube, or wherever you’re listening. We’ve got more conversations coming your way, and you won’t want to miss them.

Thanks for tuning in.