AI can help a regenerative medicine clinic follow up with patients by taking four jobs off the front desk. It can draft the message, time it, route it to the right person, and remind. What it can’t do is decide what’s true, or what your practice is allowed to say, and so a person still reads it first.
TLDR: Four follow-up jobs can come off your front desk and go to AI, but it still can’t decide what’s true or what your practice may say. So a person reads anything about a claim, a price or a clinical question before it sends. And the same rules apply when you follow up with patients by text or email. Marketing texts need written consent first, and email needs a way to opt out.
Important Note
This article is for educational purposes only and does not constitute legal, medical, or regulatory advice. Marketing strategies discussed should be reviewed by qualified legal counsel before implementation, particularly regarding FDA, FTC, and state-specific advertising regulations. Regen Portal is a marketing and consulting company, not a law firm.
One thing that I see in almost every practice I sit down with: a consult that doesn’t close on the day is rarely contacted again. The person came in, they didn’t schedule, and most of the time that’s the end of it.
Why? The front desk is the bottleneck, and follow-up is the first thing to fall off. Nobody planned it that way. It just falls off.
So can AI help you follow up with patients? It can. There are four jobs it can take off your team, and there’s one it can’t touch. I’ll walk you through each one, the rules that still apply, and how to start small.
The Short Answer: Four Jobs AI Can Do and One It Can’t
AI can do four follow-up jobs: draft the message, time it, route it to the right person, and remind. It can’t decide what’s true, and it can’t decide what your practice is allowed to say, so that job stays with a person.
So can it do all of your follow-up? The short answer is: it depends on which part. AI can handle tasks, but it can’t replace human judgment, and so I ask the same question about each job. Is this a task, or is this a judgment call?
This is how the four jobs split.
| Job | What AI Does | What a Person Does |
|---|---|---|
| Draft | Writes a first version from approved wording | Reads it and approves it |
| Time | Sends on the schedule you set | Sets the schedule and stop rules |
| Route | Sends each reply to the right person | Answers the question |
| Remind | Flags open replies and next steps | Owns the list |
What this means for your practice: Look at the right-hand column and ask if somebody on your team owns each of those today. If not, fix that first.
Where Does the Follow-Up Break in a Real Practice?
It breaks at the front desk, and it breaks early. The admin or the receptionist is usually the one tasked with the outreach, and it’s rarely automated. It can be in nearly every case, and it should be.
How early? In the practices we’ve looked at, a new inquiry can wait a few minutes for its first touch, or it can wait a few days. Minutes, or days. And that’s only the first touch, so what do you think happens to the second one?
Another thing that I see in almost every practice: leads land in three systems that don’t talk to each other. So who has the one list of people waiting on a reply?
I don’t blame the front desk. Follow-up is one more job on top of everything else they do.
The Four Jobs: Draft, Time, Route and Remind
Each of the four jobs is a task. Drafting gives you a first version, and the other three get the message out and get the reply to a person.
Start with the draft. You give the tool your approved wording, and it writes a first version of each message. Each message has one job inside a sequence built email by email, and so I’d keep each draft to the one thing that person asked about.
And notice I didn’t say paste the chart in. I’ll get to that.
Then timing. Who decides when the second message goes out? You do. You set the schedule once and then the tool keeps it, because it doesn’t get busy and it doesn’t forget. You set the stop rules too, like when the person schedules or asks you to stop.
Third job, routing. A patient writes back with a question, and where does that reply go? AI can read it and send it to the right person. So a clinical question goes to the provider, and a price question goes to whoever owns the price conversation. The tool routes. That’s all it does here. It routes.
And then reminding. It reminds the patient about the visit, and then it reminds your staff that a reply is still open. Who chases the open reply at four in the afternoon? The tool can.
The Human Gate: What a Person Reads Before Anything Sends
A person reads anything that touches a claim, a price or a clinical question before it goes out. The tool can write it, but somebody on your team has to say yes first.
Why so strict? Because AI is here to please you. You ask it for a warm follow-up message and it gives you one, and if a bigger claim sounds warmer, it can write the bigger claim. It doesn’t know what your practice is allowed to say, because it only knows what’s published.
Every day I come across ChatGPT output on a clinic’s blog, website, forms and socials that nobody reviewed before it went up. Every day. So why would a follow-up message be any different?
I’ve also seen after-hours replies make claims or quote prices they shouldn’t have. It’s rare, because AI-driven after-hours handling is still lightly used in this industry. But I’d build the gate now, before it becomes common.
So what does the person at the gate look for?
- Does the message say, or hint at, what a therapy will do for that person? Pull it.
- Does it quote a number or a special offer? The price conversation is already improvised in almost every practice I sit down with. A tool shouldn’t improvise it too.
- Did the patient ask something only a provider should answer? Then it goes to the provider.
Everything else is logistics, like when the visit is and where to park. That can run on its own once you’ve approved the template.
Which Rules Apply When You Follow Up With Patients by Text or Email?
No rule goes away because a tool hits send. HIPAA comes first, if your practice is a HIPAA covered entity. Then written consent for marketing texts, and the opt-out rules for email.
I’ll tell you what each page says, and counsel decides how it applies to your practice.
Let’s say your practice is a HIPAA covered entity. HHS has a page on business associates, last reviewed in July 2026. A business associate is a person or organization that does a job for you involving protected health information. You may share that information with them if you have “a contract or other written arrangement.” That’s the business associate agreement, or BAA.
So does the AI tool that holds your patient messages need one? That’s the question I’d take to counsel before a single name goes in. The same question comes up with tracking pixels, forms and chat tools, and so the AI tool is one more name on that list.
Still under HIPAA, and still only for a covered entity, there’s the marketing rule. HHS has a definition of marketing: “a communication about a product or service that encourages recipients of the communication to purchase or use the product or service.”
With limited exceptions, the rule requires a person’s written authorization before their health information is used for marketing. So read your reactivation message one more time, and ask if it encourages someone to buy or use a service. That’s a question for counsel.
Now texts. The federal rule on marketing texts says marketing texts and automated calls need written consent first. People also need a working way to stop them. From November 25, 2026, an FCC order lets a sender name one way to opt out, such as replying STOP, if every message says so clearly. A STOP reply to a marketing text ends all marketing texts from that sender. Counsel has the details.
And email. The FTC’s CAN-SPAM compliance guide lists what the law asks for, like no deceptive subject lines. Tell people how to opt out, and honor an opt-out request within 10 business days. There’s one more line in that guide I’d put on the wall: “Monitor what others are doing on your behalf.”
So if somebody else sends email for you, you still have to watch what goes out.
The Written Standard: What Staff May Put Into Which Tool
Write two things down before anyone on the team uses a tool to follow up with patients. The first is a prompt library, so everybody starts from the same approved wording. The second is a standard for what staff may put into which tool.
Why write it down? In almost every practice I sit down with, staff are already using consumer AI tools, unsupervised. So who told them what’s okay to type in?
When we do AI and ops consulting with a practice, the team trains in live working sessions on the practice’s own cases, and they leave with both in writing.
What does the standard look like? It names each tool, and it says what can go in and what never goes in. Most marketing prompts hold no patient data. The danger is the patient-specific prompt, with a name or a condition in it. I’d put those on the never list until counsel tells you otherwise.
And all of that work ends in a documented runbook, and a retraining session when staff turn over.
How to Start With One Sequence
Start with one sequence, and I’d make it the consult that didn’t schedule. Get that one running with a person at the gate, and then add a second one.
When we work on patient communication with a practice, it comes down to four kinds of sequences: reminders, preparation, follow-up and reactivation. I wouldn’t build all four at once.
So why start with the consult? Go back to the top of this article, because that’s where it breaks. The patient is making a big decision at the consult, and so some of them leave without a date. One sequence is enough to follow up with patients like that.
This is the order I’d do it in.
- Pick the one group: consults with no visit on the calendar.
- Write the messages yourself, or approve every draft.
- Set the timing and the stop rules.
- Name who gets each kind of reply.
- Take the consent and opt-out questions to counsel, and the tool agreement if you’re a HIPAA covered entity. Do it before a name goes in.
Then read what went out, every week at first. Once that one runs without you chasing it, you can decide which workflow to fix next.
So it all goes back to one question. Who on your team reads it before it sends?
Take a Practice That Starts With One Sequence
Take a practice with three providers, an office manager who also covers the front desk, and a regenerative service line that’s two years old. This scenario is illustrative and does not describe any Regen Portal client.
The Challenge: People come in for a consult and leave without a date. The office manager means to follow up with patients who leave that way, and then the phones start and the list never gets worked.
The Approach: The owner picks that one group. The tool drafts a short set of messages from wording one provider already approved, and then it sends them on a schedule the owner set. A reply with a question goes to a named person, and the tool reminds that person until it’s handled.
The Compliance Check: Before anything sends, that provider reads every draft, and anything the gate is there to catch goes to a person. Counsel looks at the consent wording for texts and the opt-out in the emails. And since this practice is a HIPAA covered entity, counsel also decides whether the tool needs a business associate agreement.
What Changed: Every consult without a date gets contacted, and every reply has an owner. Did more people schedule? I wouldn’t promise that to anybody. What changed is that the office manager stopped carrying the list in their head.
Frequently Asked Questions
Can AI Text My Patients for Me?
Yes, it can send the texts, once the consent question is answered. If your practice is a HIPAA covered entity, the question about the tool agreement comes first too. Marketing texts need written consent first, and people need a working way to stop them. Counsel confirms the details before you turn anything on.
Is It a HIPAA Problem to Use AI for Follow-Up?
It can be, if your practice is a HIPAA covered entity and patient information goes into a tool with no agreement behind it. The question to ask is whether that tool needs a business associate agreement, and I’d ask counsel first.
What Should AI Never Say to a Patient?
It should never say what a therapy will do or what it costs, and it should never answer a clinical question. The tool can route the question, and that’s where its job ends.
How Soon Should a Clinic Follow Up With Patients?
As soon as you can do it the same way every time. I don’t have a magic number for you, so pick a window your team can keep on the busiest day of the week.
Will AI Replace My Front Desk?
No. AI takes tasks off the front desk so the person there has time to talk to people. Somebody still has to answer the question the tool routed, and that’s your front desk.
Key Takeaways
- AI can help you follow up with patients by doing four jobs: draft, time, route and remind.
- It can’t decide what’s true or what your practice is allowed to say.
- A person reads anything about a claim or a price before it sends, and clinical questions go to a provider.
- The same rules apply when a tool sends: HIPAA if you’re a covered entity, consent for texts, opt-out for email.
- Write down what staff may put into which tool, then start with one sequence.
Let’s Talk Strategy
I sit down with a practice, I look at where the follow-up breaks, and we build one sequence at a time.
Do you already have a message your team uses to follow up with patients? Or one a tool wrote for you? Email it over and I’ll read it.
PS: If somebody is setting up AI follow-up for you, I’d find out who reads each message before it sends and where a reply goes. Ask the questions.
Email: hello@regenportal.com
YouTube: Subscribe to the channel
About Regen Portal
Regen Portal is a marketing and consulting company built for the regenerative medicine industry. We handle practice consulting, marketing compliance, search engine optimization, content creation, social media management, paid advertising, website development, branding, and AI and ops consulting for clinics, manufacturers, distributors, and independent providers. For more on me and how we work, visit Regen Portal and contact us.
About the Author
Oscar Tellez is the founder of Regen Portal, a marketing and consulting company built for the regenerative medicine industry. He has been in this industry since 2015. He started in medical sales and has worked with PRP since 2016. He has reached hundreds of practices and over 200 doctors through sales, teaching and marketing. He holds a B.S. in Exercise Science and Health Promotion from Florida Atlantic University.
