Shipping an enterprise site on a tenth of the expected budget.

NCR’s hospitality division had a product launch scheduled for early 2021 and needed the ncr.com/restaurants section rebuilt around it. My CMO secured $72,000 to fund the work. On paper the number looked workable. Then the CEO recommended a specific AEM agency, and what had been a financial constraint became a political one.

AEM implementations carry two costs that compound: specialized development talent and extended timelines. My experience with AEM came from direct exposure to how those projects actually run — and it told me $72,000 wouldn’t get a site shipped. Finding an agency willing to scope at that number would be difficult. Getting one to deliver at it was a different problem.

The real constraint topology was unforgiving: CEO preference pointed to a specific agency; that agency operated at a cost that would consume the budget before development finished; the launch date was fixed; and a delay caused by a failed vendor decision — with a CEO-backed agency already named — would have a clear owner, and that owner was me.

The path didn’t come from a framework. It came from platform knowledge built through direct implementation work. I’d been running Webflow on microsites and had worked far enough into it to understand what it could do at enterprise scale. Webflow had just released its Enterprise product, and that changed the budget math: the license cost left room for a meaningful agency contract alongside it.

Rather than make that call alone and defend it later, I proposed an accelerated RFP. The design of it mattered: I wasn’t running a standard vendor selection, I was building a data instrument that could surface the gap between the CEO’s preferred agency and market reality without framing it as a personal challenge. Working with Procurement and Legal, we compressed a six-to-eight-week process into two. Seven agencies completed the RFP.

The results pointed to Edgar Allan, a small Webflow agency with genuine technical depth. Their proposal fit the budget, and their answer to the domain-architecture problem — a reverse-proxy configuration serving Webflow under ncr.com’s actual URL path — was the only one that held up under scrutiny. That gave me the data to move the internal conversation. IT defaulted to AEM, a conservative position with real logic behind it; I worked the objection directly. Webflow could run inside NCR’s AWS infrastructure and satisfy security and domain requirements.

Edgar Allan built the site using a reverse-proxy configuration alongside Webflow’s static page builder. Webflow served content under ncr.com/restaurants as a URL path routed through NCR’s AWS architecture. Visitors stayed inside the NCR domain throughout. IT kept control of the security boundary. Marketing got a brand-aligned site that launched on time, within budget, and ahead of the hospitality product launch.

The architecture was sound because it was built around what NCR’s infrastructure actually required — not what a standard Webflow deployment would have looked like, and not what AEM would have demanded.

The project didn’t close when the site launched. The proof of concept established here became the direct input into NCR’s company-wide CMS evaluation two years on — the platform decision that carried through the NCR Atleos spin-off. One site section, one product launch, one narrow implementation: the seed of an enterprise-level platform shift.