I spent over four years as VP of Marketing for a multi-brand, multi-location consumer company, scaling revenue from $20M to $60M across multiple locations. That experience is where a lot of my read on this comes from, not just theory.
The appeal of off-the-shelf, and where it runs out
Every point-of-sale system, every scheduling tool, every reporting platform you already use is racing to add AI features right now — and a lot of them are genuinely useful. Before I ever propose a custom build for a client, I check whether their existing stack already covers it, because it often does, and it's usually cheaper to turn on a feature you're already paying for than to build something from scratch.
But generic AI features are, by design, built for the median customer of that software company — not for your specific footprint, your specific locations, or the specific way information moves (or doesn't) between your sites and your central team.
Where the real gap shows up
For multi-location brands specifically, the pain is almost always in coordination: how does a decision made at headquarters actually reach every location consistently? How do you compare performance across sites without someone manually assembling a report? How do you catch the location that's quietly underperforming before it shows up in a quarterly review instead of a weekly one? Off-the-shelf tools rarely solve this well, because it's specific to your org chart and your locations, not a generic problem every customer of a platform shares.
What a custom system actually buys you
A purpose-built system is designed around your specific structure — your specific locations, your specific reporting chain, your specific definition of "underperforming." That specificity is exactly what off-the-shelf tools can't offer, because it would mean building something different for every customer, which isn't how SaaS economics work.
The honest tradeoff
Custom costs more upfront than a software subscription, and it takes real discovery work to get right. It's worth it when the coordination problem across your locations is costing you real money every month — not when a generic feature would genuinely solve it. The right answer starts with a clear-eyed look at what you already have before deciding which one you need.
Before I ever propose a custom build, I check what your existing tools already do. That review is part of the AI Opportunity Assessment — worth doing before spending on anything custom.