Should I hire a setter or let AI follow up with my leads?

Why this is the question now

Six of the 58 sales calls in our backlog since June asked some version of "can AI do this instead of hiring?" Four more, all from solo operators, asked it the way it actually feels: "I got three leads today, I only have time to call one." The ads were working. The leads were arriving. Nobody was free to pick up the phone, because the owner was in a treatment room with a patient who had already paid.

That is the real comparison. Not AI versus a person in the abstract, but which one solves the first hour after a lead comes in, and what each one costs you in money and attention.

What a lead needs in the first hour

A new lead is a stranger who just told you they have a problem. They filled in a form at lunch, or at 10pm after the kids went down. In that moment they are also looking at two other clinics. Whoever answers first, by text, usually gets the conversation, and the one who answers tomorrow gets a voicemail that never comes back. We have watched this in client accounts for years: a lead that sits for an hour has often already found someone else.

Each of those leads cost real money. Across the Campfire Benchmark, the three clinic Meta ad accounts that track a booked appointment (the pixel or CAPI Schedule event) spent $17,946.51 for 136 bookings, which is $131.96 per booked appointment, pooled. Every lead that goes cold because nobody answered is a piece of that spend thrown away. So the follow-up system is not an add-on to the ads. It is the part that decides whether the ads were worth running.

What a setter does well, and where it costs you

A good setter is a good conversation. They hear hesitation in a voice, handle "how much is it" without flinching, and can tell a tire-kicker from someone who is scared and needs reassuring. For a clinic selling a plan of care, that human read on the second or third contact closes patients that a text thread never will.

The costs are the ones owners underestimate. A setter is a fixed monthly expense whether four leads come in or forty. They need a script, a CRM to work from, training on your offer, and someone to check their notes, which is you. They work set hours, so the 10pm lead still waits until morning. And when they leave, the follow-up process leaves with them unless it was written down somewhere. On calls, the owners who have tried this usually describe it as hiring a person to do a job they never fully defined.

What AI follow-up does well, and where it stops

Software is good at the boring, time-critical part. A lead fills in the form and gets a text within a minute, at any hour. The text asks two or three qualifying questions, so the ones with the wrong insurance, the wrong age, or a problem you don't treat sort themselves out before anyone's time is spent. The ones who fit get a calendar link and book a discovery call themselves. If they go quiet, a two-week sequence of texts, emails and a voicemail drop keeps the door open without anyone remembering to. You get a notification for every form fill and a short morning list of who to follow up with, so nothing depends on you checking.

Where it stops is the conversation that needs a person. AI should handle the basics, after-hours replies, intake and speed to lead, and a human should convert. If a lead replies "I had a bad experience with PT before," that is not a job for a bot, and we don't pretend it is. The AI's job is to make sure that reply arrives in front of a human who is free to answer it.

How to decide

Ask these in order:

• Are leads going cold before anyone replies? If yes, install AI follow-up first. A setter cannot fix speed; they still sleep.
• Do you get more qualified leads than you can hold real conversations with? If yes, that is the moment a setter pays for themselves, and the AI has already done their screening.
• Is your lead volume under a handful a week? Then neither a setter nor a complicated system is the priority. Fix the ads and the offer, and let the automation hold the leads you do get.
• Can you describe the follow-up process in writing? If not, build it in software first. Once it works, you can hand it to a person and they inherit the process instead of inventing one.

The setup we install

For most clinics we build the hybrid: a lead form whose last step is "book a discovery call," so leads self-schedule; the AI sequence for the ones who don't; and a two-minute human pre-qualification call before the discovery visit. That call can be your front desk, you between patients, or, if you are solo, someone on our team during the build. The person shows up where the money is decided, and the software does everything before that. Our article on speed to lead for clinics covers the timing in detail, and the neutral guide to speed to lead on Clinic MarketingOS has the industry numbers behind it.