Most companies don’t design how they work. They inherit it.

There’s a concept in archaeology called a tell. It looks like a natural hill but isn’t. It’s the remains of one settlement built on top of another, generation after generation, each flattening the last and building on the rubble. The surface looks like solid ground. It’s layers of decisions nobody alive remembers making.

Most business processes and systems are also tells.

The approval chain that takes four signatures? The report three departments quietly depend on? The handoff between sales and ops everyone complains about but no one really “owns”? Almost none of it was designed. It accumulated. A reorg here, a departed employee there, a system bought in 2012, a workaround that outlived its problem. Software running rules nobody remembers writing. Layer upon layer, until the surface looks permanent.

Then someone decides it’s time to fix it with whatever magic beans are handy or in vogue.

But here’s the part companies keep missing: digitizing an inherited process, automating it, or pointing AI at it doesn’t excavate the tell. It risks paving over it.

The same shape at every size

I’ve seen this play out at both ends of the size spectrum.

A small professional services firm had accumulated a stack of tools over the years. A homemade CRM. An accounting package. A scheduling tool. File-sharing. A password manager. A project board. Getting one prospect into the pipeline meant a chain of manual steps: enter them here, hope they fill out the form, add them to a list, create a folder, chase them for documents, stash their credentials in three places. A savvy user tried wiring it together with an integration platform, but it only automated part of what was needed. Now worse, not better.

Nobody had ever looked at the whole workflow and asked how it should flow. So people became the integration. Staff carried in their heads who was where in the pipeline, what everyone was waiting on, and what needed to be entered where to keep it all together.

A very large specialty contractor had the same shape at a bigger scale. Estimating in spreadsheets. Contracts in Word. An off-the-shelf CRM. A best-in-class tool for project management. A separate accounting system. Another for capacity scheduling. Email as the de facto glue between all of it. They once handed me a spreadsheet listing every place the same piece of information gets keyed in by hand more than once. Dozens of items long, most entered into two to five systems, once when the record was created and again every time it changed. None of the systems were truly connected.

Again, this was not a designed “system of systems,” but one that grew organically, fueled by well-intentioned humans who became the connective tissue that bound it all together.

Neither company made a bad decision on purpose. Each tool was bought by a committee or a capable user solving a real problem in the moment. But no one owned or designed the overall flow. The tools accumulated, and people quietly became the thing holding it together.

The catch

When people are the integration, you can’t scale the work without scaling the number of people holding it in their heads. And when it finally hurts enough, the instinct is to buy one more tool or automate what’s already there. But automating a process nobody designed just makes an accidental process run faster. Automate a bad one and you’ve made it automatically bad.

This is the quiet cost of digitizing without redesign. Technology preserves processes. Once a way of working is encoded in software, it gets harder to change, not easier: now there’s a system of record, a training cost, and an integration to worry about. Yesterday’s workaround is load-bearing, and swapping it out is tricky business.

The planner’s move

Here’s where the metaphor usually stops. But it’s also where it gets interesting.

Building on the tell isn’t the mistake. It’s how cities work. Dig under London and you hit the Victorian city, then the medieval one, then Roman Londinium, each street tracing a line someone drew for reasons long gone. Sometimes a city builds on top of what’s there and never looks back. Sometimes it does something harder: it keeps a historic structure and designs something new into it, or guts a landmark to its shell and rebuilds inside.

And sometimes the only honest move is to clear the site, at real cost and disruption, to make room for what needs to go there.

What separates a great city from a congested one isn’t whether it has a tell. They all do. It’s whether anyone is planning. A good planner starts with a vision of what good looks like, then works back into the site as it actually is, deciding deliberately what to preserve, what to remodel, and what to clear.

That’s the move most businesses skip. Not the software. Not the automation. The design. The willingness to look at the workflow with fresh eyes, decide what good looks like, and only then figure out what to keep, what to connect, and what to replace, even when replacing it hurts.

Digitizing the inheritance is the easy move. It feels like progress, and from the outside it even looks like progress. But if the underlying process wasn’t intentional, faster and shinier won’t make it more so. It’ll just make it more permanent.

Look before you leap

None of this is an argument against new software, automation, or AI. Those are often exactly the right answer. It’s an argument against reaching for them before you’ve decided what good looks like, before you know whether you’re building over, remodeling what’s there, or clearing the site.

Consider asking yourself a few questions that will help you think more like a planner, and less like a “paver”:

• Did we design this process, or inherit it?
• If we were starting today, would we build it this way?
• Do we know what truly better would look like, before we choose what to keep, replace, or add?
• What are we about to make something sub-par even more permanent by enshrining it in another layer of technology?

To be clear, I’m not suggesting that you have to excavate the whole mound. That’s rarely the right answer. But you should know you’re standing on one, and decide, in that context, and on purpose, what to build next.