Reframing an app build as a discoverability problem.

The Allpoint network is one of NCR Atleos’s flagship assets — 55,000 surcharge-free ATMs across the United States, white-labeled and co-branded for financial institutions including Chime, and embedded everywhere from gas-station chains to big-box stores. The model is a network-effects play: more locations attract more FI partners, which justifies more placements, which draws more cardholder traffic.

None of the customer-facing infrastructure was working as a system. There was a finder application, but it sat on the website behind enough friction that most people never found it. There was a partner portal. There were several ancillary properties that had grown up around the core product with no clear owner. I was brought into a meeting with the GM of the network and the CMO I reported to, and handed a directive: figure out what we have, figure out what’s broken, and fix it.

The harder part wasn’t technical. The company had never fully worked out what the Allpoint digital presence was trying to accomplish — assumptions about the app, assumptions about the portal, budget conversations happening without a clear picture of the solution. A presenting problem, the app needs work, but no diagnosis. And three audiences pulled in different directions: retailers wanted footfall, financial institutions needed their customers to find cash access, and consumers wanted something that worked the way they already searched.

The initial ask was to fix the app. That assumption needed pressure-testing before any build work began. I spent about two weeks in discovery — not a formal project, just direct conversations. I reached out to the person who ran ATM operations, because he had the closest view of how the network actually functioned day-to-day; he connected me to the team that owned the locator application and its database, and a few conversations with IT filled in the architecture.

What emerged was that the locator database was well-structured. It stored location coordinates, branding attributes, machine type, and had the downstream capacity to surface uptime data. The data was there. The problem was how it was being surfaced — or not surfaced.

So the real question became: where are people actually looking for ATMs? The answer was obvious once it was asked plainly. People open Google Maps or Apple Maps, type “ATM near me,” and tap the first result. Nobody was visiting the Allpoint website to use a proprietary finder, and a better app was never going to change that. The path forward was to meet people where they already were — local SEO, syndicating all 48,000 NCR Atleos-owned ATM locations into Google Business Profile and Apple Maps through API-based listing management. Common practice in retail and franchise; it had simply never been applied to this network.

The architecture was straightforward by design. The locator database was already the system of record. The work was a reliable translation layer between how that data was stored and what the Maps APIs required — coordinate normalization, attribute mapping, schema compliance, and ongoing sync so listings stayed current as ATMs were added, moved, or removed.

I brought in Render SEO to execute and manage the listing infrastructure. It mattered that procurement already had a relationship with them — in a regulated environment, legal review alone could run sixty to ninety days, and the existing contract path compressed that to roughly two weeks. Funding came from an unusual place: I negotiated to redirect part of the budget held by a team that had arrived through an acquired crypto-ATM company and was sitting on money without a clear deployment path. That funded the work without touching the core marketing budget, and gave that team a credible project to point to.

Beyond discoverability, there was a clear path to operational intelligence: ATM uptime data fed back into the listings to suppress or deprioritize machines that were down, so a cardholder is never sent to a dead machine — tied directly to the footfall promise retailers cared about. It wasn’t in the initial scope, but it was on the roadmap. Then the company announced it was being acquired, and almost every active initiative was paused. The foundation was complete: vendor in place, budget secured, pipeline scoped, procurement cleared.