Where a banking product meets the member it might surprise

I treat a deposit account's fees as part of its interface. When a customer pays a fee she did not expect, I look first at what the app showed her at the moment of the transaction: the available balance, pending card holds, held deposit funds and scheduled payments. At Corriston Bank I redesigned overdraft and returned-item fees around that evidence, adding a projected end-of-day balance, a window until the end of the next business day to cure an overdraft, a US$50 buffer and the removal of returned-item fees altogether. I work with finance on the revenue effect, with compliance on disclosures and on the opt-in that federal rules require before overdraft fees apply to everyday debit card and ATM transactions, and with complaints teams on the cases. I do not set prices alone. I make sure the people who do can see which fees come from customer choices and which come from the institution's own presentation.

On a platform used by 74 credit unions, accessibility depends as much on what clients are permitted to configure as on how engineers build the components. I own Lanternfield's design system, its limits on customisation and the testing programme that checks whether members can complete real tasks. We treat version 2.1 of the Web Content Accessibility Guidelines at level AA as a minimum, and we measure eight ordinary banking tasks with a paid panel of 40 members who use screen readers, magnification, switch access or voice control, alongside older members and members with cognitive disabilities. I report completion rates to our board committee by group and by task, and I decide with our accessibility lead when a client's request for an exception can be granted. We have granted four. A defect that blocks one of the eight tasks stops a release, which has delayed two releases since 2023.

I lead the product side of how third parties reach our members' account data. Until 2025 most of that access came from aggregators signing in with members' own stored passwords, which no credit union could attribute or limit. We built token connections with scoped permissions and a Connected apps page listing every active connection under the name of the app the member actually used, and we set a date for each aggregator after which password sign-ins were refused. I work with our chief technology officer on the interfaces, with our security team on account takeover and with client credit unions on what members are told. The federal rule on consumer data access is not settled, so I describe our approach as practice rather than compliance and have designed it to adjust when the rule does.

Credit unions serve many businesses that are too small for a bank's commercial team and too important to their owners to be treated as consumer accounts with a different logo. I lead the business banking module our clients offer to about 52,000 business members. Our research with owners kept returning to one question, whether the business could meet payroll, rent and suppliers over the coming weeks, so we built a thirteen-week cash-flow view from account history, scheduled payments and receipts the owner expects. I also designed dual approval for payments, delegated users with their own limits and a referral to the credit union's business lenders when a projected shortfall appears. I decline features that duplicate the accounting software our business members already use, because a credit union's advantage is the account and the relationship with its lender, not the ledger.

Every hold, warning and extra step in a banking app costs honest members time, and I design that friction as deliberately as any feature. When our fraud team asks for a new control, I ask what the member will see, what she can do next and how long any delay will last, and we test the wording with members before release. For a large first transfer to a new recipient, the shared app now asks two plain questions about how the member came to know the recipient instead of showing a generic warning, and it explains any delay with a time rather than an apology. We measure abandonment by honest members alongside losses prevented, because a control that stops a little fraud while sending many members to a branch has costs the fraud report never shows. Dispute screens follow the same rule: the first page tells a member what will happen and when.

Lanternfield's clients are also its owners, and seventy-four credit unions rarely want the same thing in the same quarter. I run product governance so that disagreement is visible and decided rather than settled by whoever lobbies hardest. Each quarter our client advisory council scores proposed work against published criteria: member harm, regulatory need, breadth of demand and cost. I publish a decision log to every client, including the requests we declined and the reasons. When a credit union wants to bring a fintech partner into the shared app, my team runs product and data due diligence with our third-party risk team, covering fees charged to members, data use, complaint handling and how the partner would leave. After several years of supervisory action against banks whose fintech partnerships outran their oversight, our clients rely on that review, and I make sure it is able to say no.