Before You Add AI to Your Healthcare Practice, Fix the Process

I spend a fair amount of time talking with business owners about AI.

The conversation often starts with tools.

“Which AI should we use?”

• Can we get an AI receptionist?
• Can AI automate our follow-up?
• Should we add a chatbot?

Those are reasonable questions, but they're not where I'd start.

My first question would be:

What process are we trying to improve?

Although AI can make a good process dramatically more efficient, it can also make a bad process fail much faster. AI is a multiplier.

This is how I describe it in my book:

AI is a multiplier, not a magic fix.

Suppose your practice has a poor process for handling new patient inquiries. Nobody clearly owns follow-up. Contact information gets stored in different places. Staff members aren't sure how many attempts to make. Nobody tracks whether inquiries eventually schedule.

Now imagine adding AI.

Congratulations! You may have just automated confusion.

Before automating the workflow, define the workflows, including what happens when someone contacts you, who owns the funnel, which steps require a person to intervene, and which steps could be automated. After that, measure and document how well the process is working, or not.

Then, AI becomes interesting.

Start With Friction

I think some of the best AI opportunities in healthcare practices are surprisingly unglamorous. They're repetitive tasks that consume staff time or create unnecessary friction for patients.

Consider missed calls.

Someone calls your PT office while the front desk is helping another patient, but nobody answers. Traditionally, the caller leaves a voicemail, someone listens to it later, calls back, gets voicemail themselves, and phone tag begins.

A relatively simple automated workflow could immediately send a text acknowledging the missed call and provide a link to request or schedule an appointment. That doesn't replace your front-desk team, it gives them help. It also gives the prospective patient an immediate next step.

Other examples include appointment reminders, routine follow-up, intake workflows, reactivation messaging, and handling certain common questions.

These are the types of practical applications I focus on in the AI Leverage System.

The Goal Isn't to Use the Most AI

There's a strange competition happening in business right now around who is “using AI.”

I don't think that's a useful metric.

If you have twelve AI tools and none of them meaningfully improve your practice, I'm not impressed.

If you have one automated workflow that saves your staff five hours a week or prevents qualified prospective patients from disappearing after missed calls, that's interesting.

Technology should earn its place.

Ask:

• Does this save meaningful staff time?
• Does it improve responsiveness?
• Does it remove friction for patients?
• Does it reduce unnecessary administrative work?
• Does it help people make better decisions?
• Can we measure the result?

If the answer is no, adding AI because everyone is talking about AI isn't much of a strategy.

Keep Humans Where Humans Matter

Healthcare creates an important boundary. Patients want convenience, but they also deeply care about trust. Those aren't competing ideas.

Use technology to handle repetitive friction so people have more time for interactions that benefit from being human. An automated system can remind someone about an appointment and a person can help them when their situation is complicated. AI can send an immediate response after a missed call, but your staff can have the actual conversation. An AI receptionist may handle an after-hours scheduling request, while your provider still builds the clinical relationship.

I don't view that as replacing the human experience. Done properly, it's protecting time for it.

A Good First AI Project Should Be Boring

This may be my least exciting AI advice:

Start small.

Don't begin by trying to “AI-enable the practice.”

Pick one annoying, repetitive process.

Map what happens today.

Find the friction.

Determine what should remain human.

Automate a piece of it.

Measure the result.

Then decide whether to expand.

Your first AI project doesn't need to transform the healthcare industry.

It could simply mean fewer missed opportunities and two hours a week returned to your front desk. That's a win.

Strategy First, AI Second

AI technology will continue changing rapidly. Individual tools will come and go, but the more durable skill for a practice owner is learning how to identify processes where technology creates genuine leverage.

That's why I wouldn't start with:

“What can AI do?”

I'd start with:

“Where are we wasting time or creating unnecessary friction?”

Find that problem first.

Then decide whether AI is the right tool to solve it.

If you're trying to figure out where AI could create practical value in your healthcare business, I help practices evaluate the strategy, processes and technology together. You can book a discovery meeting with me to discuss where it makes sense to start.