Healthcare Doesn't Have a Data Problem. It Has an Action Problem.

Healthcare organizations aren't short on data. We have claims data, utilization data, member engagement data, CRM data, call center data, provider data, client meeting notes, emails, implementation trackers, satisfaction surveys, and financial reports.

We have dashboards for our dashboards.

And yet, when an employer client asks a relatively simple question like, “What should we be doing differently?” the answer can be surprisingly difficult to produce.

That's because I don't think healthcare has a data problem. We have a fragmentation problem. And increasingly, an action problem.

The Information We Need Is Everywhere

Imagine you're an executive responsible for employer clients at a health plan, benefits company, or healthcare organization. Your CRM tells you when the client renews. Your claims system tells you what's being utilized. Your engagement platform tells you who's interacting with your programs. Your implementation team has a project tracker, and your account team has meeting notes.

Meanwhile, your client may have emailed three different people about concerns. Somewhere in Gemini meeting notes, Gong, Fireflies, or someone's notebook is an important comment like, “Our CFO is questioning whether employees are getting enough value from this program.”

Every piece of that information matters. The problem is that no single system tells you the whole story, so we rely on people to connect it.

And the more clients, systems, and data we add, the harder that becomes.

That's Where AI Gets Interesting

For me, one of the most compelling applications of AI in healthcare isn't generating another report or writing another email. It's connecting information that already exists and helping us determine what to do next.

Imagine that before an employer meeting, an executive receives a short briefing. It tells her that increasing utilization and demonstrating ROI are the client's current priorities. It reminds her that the client said employees don't understand how to access the program. It surfaces the team's commitment to provide a new engagement recommendation. It also identifies that utilization has improved slightly, but one employee population continues to significantly underutilize the benefit.

Then AI adds another piece of context: ROI has now been raised in two consecutive meetings, and renewal is approaching.

The recommendation isn't another utilization report. It's to walk into the next meeting prepared with a targeted engagement strategy and a point of view about what should happen next.

None of those individual pieces of information are revolutionary.

Connecting them is.

From Systems of Record to Systems of Action

For years, healthcare organizations have invested heavily in systems of record. We got better at storing information. Then we invested in business intelligence, and we got better at visualizing information.

I think the next opportunity is getting better at acting on information.

Instead of simply asking AI to summarize a meeting or draft an email, imagine being able to ask questions across previously disconnected sources.

What commitments have we made to this client? Which employer clients have raised the same concern multiple times? Which clients are approaching renewal with unresolved issues? Where is utilization declining? What did we tell the client we would do about it? What should the account executive know before walking into tomorrow's meeting?

That's a much more interesting use of AI.

It's also a fundamentally different way of thinking about the technology. The value isn't necessarily in creating more content. It's in helping people make sense of the information the organization has already created.

The Technology Isn't the Hardest Part

Of course, connecting healthcare data isn't as simple as plugging every system into an AI tool. Privacy, security, permissions, data governance, accuracy, and appropriate human oversight all matter enormously.

And more data doesn't automatically produce better decisions.

The goal shouldn't be, “Let's give AI access to everything.” The better question is, “What information does this person need to make a better decision, and how can we responsibly bring it together?”

That distinction matters.

Organizations don't necessarily need to start with a massive AI transformation. They can start with a specific decision, workflow, or client interaction and work backward. What does the person need to know? Where does that information live today? What requires human judgment? Where could AI remove the manual work of finding, connecting, and summarizing it?

Those questions can lead to much more practical AI use cases.

The Human Still Makes the Decision

This is also why I don't think the most valuable AI applications will remove healthcare executives from the process. They'll make them better prepared for it.

AI might identify a pattern, surface a risk, remind us of a commitment, or recommend a next action. But an experienced healthcare leader understands the relationship, the nuance, the history, and the consequences of the decision.

AI connects the dots. People decide what the picture means.

That's the Opportunity

I started my career in healthcare because I wanted to help people. Over time, I've realized that helping people at scale often means fixing what happens between systems, teams, and processes.

That's where so much friction lives.

It's also why I'm excited about this next phase of AI in healthcare. We don't necessarily need another dashboard, another report, or another place to store information. We need better ways to connect what we already know to what we should do next.

Healthcare has plenty of data. The opportunity is turning fragmented information into action.