Why the Best AI Doesn’t Replace People — It Gives Them Their Day Back

Every founder has a moment when they have to decide whether they’re solving the obvious problem or the real one.

When my company first started working with university admissions teams, people assumed we’d build a better CRM. It made sense on paper. Admissions offices rely on platforms like Slate, Salesforce, Banner, and PeopleSoft to manage applications, communications and student records.

But after spending time inside admissions offices, I realized something surprising.

Almost nobody was asking for a new CRM. 

They were asking for their day back.

The work wasn’t where we expected

The conversations that changed my perspective weren’t about artificial intelligence or digital transformation. They were about ordinary mornings.

An admissions counselor opened a shared inbox with hundreds of emails waiting. A transfer evaluator downloaded transcripts one by one. Someone manually identified document types before uploading them to another system. Advisors bounced between multiple applications just to answer a student’s question with confidence.

None of this was considered unusual.

What struck me wasn’t that these tasks existed. It was that they’d become invisible. When repetitive work becomes part of the culture, people stop questioning whether it should exist at all. 

That’s where I think many founders make a mistake. We assume the software people use every day is the problem. Often, it’s everything happening between those systems that creates the real friction.

Legacy software isn’t the enemy

Startups love replacement stories. “We’re going to build the next version” or “We’re going to replace the incumbent.”

Those narratives are compelling, but they don’t always reflect how large organizations actually operate. Universities are a good example. Their CRM connects to student information systems, financial aid platforms, learning systems, and decades of institutional processes. Replacing that foundation isn’t just expensive. It affects hundreds of workflows that depend on stability.

What admissions leaders wanted wasn’t disruption for its own sake. They wanted fewer manual steps, faster processing, and better visibility without having to rebuild everything underneath.

That changed how I think about enterprise software more broadly.

Sometimes the best innovation doesn’t replace infrastructure. It makes existing infrastructure more capable.

The hidden cost of workarounds

One lesson I’ve come to appreciate is that organizations rarely accurately measure the cost of small interruptions. Downloading a transcript takes seconds. Uploading it takes seconds. Verifying it takes seconds. Individually, those moments barely register. Across thousands of applicants, they become weeks of work. 

The same pattern appears across industries. Customer support teams copy information between systems. Finance teams reconcile spreadsheets. Healthcare staff enter the same information multiple times.

The work feels minor because it’s distributed. The cost is enormous because it’s repeated.

Intelligence belongs inside the workflow

AI has created understandable excitement, but I think we’ve framed the conversation too narrowly.

The question isn’t whether AI should replace people. The better question is whether technology can remove the repetitive work that keeps people from doing their highest-value work.

If information already exists, people shouldn’t have to re-enter it. If a document arrives, someone shouldn’t have to identify it manually before work can begin.

If context already exists somewhere in the organization, employees shouldn’t have to hunt for it across five different systems.

The most useful technology often disappears into the workflow rather than requiring people to learn an entirely new one.

A lesson I’ll carry into every company I build

One of the unexpected lessons from working with admissions teams is that success rarely looks dramatic. Nobody celebrates fewer clicks. They celebrate leaving work on time. They celebrate answering students faster. They celebrate having enough space to focus on conversations that require judgment rather than administration.

I think that’s a useful reminder for founders in any industry.

It’s tempting to chase the biggest replacement opportunity because it sounds more ambitious. But sometimes the bigger opportunity is hidden inside the workflows everyone has learned to tolerate.

The companies that create lasting value won’t always build new systems of record.

Some will simply make the existing ones finally work the way people always wished they would.