The dashboard never did anything

Here we go again. The OpenAI-Plaid deal is perking up personal financial management talks.

In exchange for connecting their financial accounts, paid ChatGPT subscribers can view their spending patterns and get money advice on things like, “help me come up with a plan to save a little bit more in the next few months.” But we have been here before in digital banking and read-only money advice didn’t take off. Some of it was a design flaw. Most people didn’t want to hunt to find budgeting features or categorize their transactions or see how little money they had in pie charts.

Even the standalone budgeting apps encountered challenges. Consider Mint, a company credited for helping make personal finance approachable. Mint was “free,” selling financial product ads instead. At times, the brand got into trouble because of this business model. Mint also ran into data connection challenges during an era of heavy screen scraping, rendering results incomplete at best. As one Reddit user posted: “I never found their budget tools useful. All it does is warn me when over budget. And often it is wrong.”

What Mint accomplished was and remains notable. It inspired people to engage with their transaction data, clarified the ways in which they were spending and encouraged setting goals. At its peak, Mint had millions of registered users. But it ultimately didn’t endure and it left a big door open: read-only transactions won’t motivate the masses to change their financial behavior. Why? It’s perceived as too much work, and for many, it was.

That kind of dashboard can show consumers they are overspending, but it won’t unsubscribe them from a subscription they no longer use or move money into an account with a higher yield. The experience is a bit akin to handing someone last month’s The New York Times and assuming they are informed. Sure, it helps. A little.

As Jordan Wright, co-founder and CEO at Atomic, wrote: “Telling a customer ‘you spent $312 on subscriptions last month’ is interesting. Canceling those subscriptions, refinancing the customer's mortgage, rebalancing the brokerage account, and rerouting direct deposits, that is what AI is about to do.”

The future of PFM ought to be about what it gets done for consumers rather than better-looking dashboards.

To be sure, that opens up big questions around trust. The industry has stumbled here before. The CFPB penalized Digit, an automated savings app owned by Opportun, after some of its users overdrafted because the app was automatically moving money from checking into savings based on what it thought they could afford. The founder Ethan Bloch, who now works at OpenAI, told me losing data connections to bank accounts and ACH timing issues were two of the reasons why it happened. The CFPB fined the fintech company $2.7 million and Digit reimbursed customers for overdraft charges (roughly $68,000).

What’s different now is payments are faster, data-sharing has grown up and LLMs are a thing. The culprits for causing flawed experiences in the past have largely improved. Certainly, companies are rethinking PFM’s next act.

Consider Revolut’s AI assistant AIR. It can pause recurring payments or activate an eSIM card when a customer is about to travel, if asked. Or, take Bank of America’s Erica, a virtual financial assistant with 3.2 billion conversations. At the Financial Brand Forum, BofA’s Head of Digital Platforms Jorge Camargo described Erica’s progression as answering questions to recommending actions to, in time, providing automated experiences, or as he said,  the ability to say: “Hey, we can do that for you. You don't have to do it yourself." But as Jorge sees it, customers won’t really welcome and adopt that autonomous experience for money movement for another five years.

While the tools have improved considerably over the years, a newer report shows that PFM utilization has dropped 9% year over year. According to the 2026 Digital Banking Performance Metrics from Cornerstone Advisors, usage declined from 24% in 2024 to 15% in 2025. The report draws on data from 89 financial institutions. The results indicate that a widow for something better remains wide open.

Roughly two decades after Mint was founded, consumers still need help with their money. Inflation is high. The price of gas is soaring. Layoffs – plenty of which are attributed to AI – are creating financial anxiety, and inertia, well, remains a big barrier into whether financial tasks get done.

If there’s a time where tools can help change at least a slice of the financial course of someone’s life, it could be now. The quality of data is getting better every day. The experiences require fewer steps from consumers to achieve something. The data-connection issues that tripped up earlier attempts have mostly matured. PFM has another opening to be useful to the masses, but only if it’s reinvented, not refreshed.

This is my first column for Atomic Insights, and I'm just getting started. In the months ahead, I'll be exploring the future of banking, fintech, and the persistent challenge at the center of both: helping people improve their financial lives.