AI setter vs human setter: should you replace your setters?

A men’s fitness coach told me on a call why he kept the setters he had. He runs two VAs on shifts from 6 a.m. to 10 p.m. On a good day one of them books five or six calls. On a bad day she books zero. But the VAs he’d hired before had to go in less than a month, so in his words, “the devil you know is better than the devil you don’t.”

Then he looked at his volume: “When I look at messages sent on a monthly basis, it’s in the multi tens of thousands. I’m like, there is no way that a human can maintain quality at that much volume.” His conclusion: “I just believe it’s a job for AI.”

I agree with him about the texting. But setters do more than text. They call leads to confirm calls, chase leads after Meta’s messaging window closes and work the story polls, and some of that still needs a person. So the useful question is which parts of the job move to AI and which stay human.

My team at OmniFusion has run AI setters for 120+ high-ticket coaches over 3+ years. Many came to us with setters already in their DMs, from one part-timer to eight or ten people.

Should you replace your setters with AI?

Replace the texting once two things are true: your DM process is proven, and more leads come in than your setters can answer well. Keep people on the phone, on the follow-ups after Meta’s 24-hour window and on the sales call.

Here’s how I’d decide, situation by situation:

Your situation	What I’d do	Why
You’re just starting and nothing in your DMs is proven yet	Do the setting yourself, or hire an experienced setter	A person can try ten things in a day. The AI needs those data points to build on
You chat with your audience in your DMs and want to keep doing it	Keep the inbox personal	The AI is built to qualify and book, not to chat for you
Your script books calls, but leads wait for the next shift, the night or Monday	Move text setting to AI	Every reply in about 2 minutes, at any hour
A reel or a lead magnet brings more DMs than your team can read	Move text setting to AI	Volume stops being a hiring problem
Your setters already call leads to set or confirm the call	Keep them on the phones	Phone setting is still a job for a person
A lead has been quiet for more than 24 hours	One trained person on follow-ups	After Meta’s window, the nudge has to come from a person
The sales call	Your closers	Closing stays human

If you’re doing $30K a month or more, your DMs already book calls and you have setters, you’re in several rows at once. That’s the usual split: the AI takes the texting, and people keep the phone, the late follow-ups and the sales call. With the AI on every DM and one trained person on the follow-ups, you never miss a DM or a follow-up again.

What does an AI setter do better than a human setter?

Texting at speed and at volume, with the same rules in every conversation. Those are the three things a setter team gets worse at as you grow.

Speed. The AI answers every message in about 2 minutes, day and night, with the timing varied on purpose so it reads like someone typing. A setter answers when they’re on shift, and covering all 24 hours with people takes at least three of them. I’d fix speed first, because everything you fix after it depends on the lead still being there. In a Harvard Business Review study of online sales leads, firms that tried to contact a lead within an hour were nearly seven times as likely to qualify it as firms that waited longer.

Volume. Human setters juggle 20 to 50 conversations at once, and the pain point a lead mentioned last week gets lost along the way. The AI holds every thread. In June 2026 it ran 20,000 conversations on Sam Johnston’s account, about 660 a day. His estimate of doing that with people: “I’d need 10 different setters. It would be a mess to track.”

Control. When you change the AI’s script, the change is live in every conversation within a day or two. With setters, you retrain each one and then check they actually use it. My team also runs a new script next to the old one on a set share of leads, 30% or 40%, until one proves it books more calls. We test openers the same way, and a winning opener usually gets two to three times the reply rate of the usual first message.

Control matters for your account too. Money questions in the DMs can get an Instagram account restricted, and a setter can go rogue and ask them anyway. Once a question is out of the AI’s script, it’s out of every conversation.

None of this works if the AI sounds like a bot. Ours is trained on your DMs, content, sales script and training material, and it’s the most human-like AI in the space. It texts in short messages, mirrors the lead’s language and remembers what they told you three weeks ago. Here’s how our AI appointment setter learns to sound like you.

Where does a human setter still win?

On the sales call, on the phone, after Meta’s messaging window and at the very start. In those four places I’d keep people.

The sales call. Your closers stay human. Once a qualified lead shows up, it’s your closer or you. What changes is how busy they get: at Kneehab Academy’s peak, the AI booked so many calls that all five closers were booked out days in advance.

The phone. If your setters call leads to set or confirm the call, keep them on it. As I told a coach whose team runs phone setters: “I’m just not convinced enough that it’s better than humans. In text, 100%, we don’t need those humans at all.”

After Meta’s messaging window. Meta’s standard messaging window for automated replies is 24 hours after the lead’s last message in most cases. Inside it, the AI follows up with text, GIFs or voice notes in your voice. After it, the nudge has to come from a person. That’s one trained person, an hour or two a day, working a follow-up board plus story views, story polls and old hot leads. We provide and train that person as part of the service. Those nudges bring 10 to 20% of ghosted conversations back, and when a lead replies, the AI picks the conversation up again.

The start. When nothing in your DMs is proven yet, a person is the better tester. You need to find out which wording, which script and which kind of lead works for your offer, and someone in your inbox can try ten different things in a day. Do it yourself, or hire a setter with real sales experience. The same goes for coaches with little DM volume. If you keep your team for now, my guide on how to manage setters and closers covers the daily numbers, the reports and the pay.

What does a setter team really cost?

$5,000 to $10,000 a month, plus commissions, before the costs that never show up on payroll. Your closers’ 10 to 12% sits on top either way, because closers stay.

That number comes from the pay structure I see most: a setter gets a fixed amount, something like $1,000 a month, plus 2 to 4% commission on the booked calls that close. It’s not written in stone, and everyone has their own approach.

Then the costs that never show up on payroll:

• Turnover. Setters quit, get fired or get good enough to earn more as closers. The founder of a setter agency that runs about 20 setter teams walked me through one month in which his team onboarded five new clients. They hired about 25 setters, let half go, hired 12 more, let half of those go and hired another six.
• Management. Someone reads the end-of-day reports, keeps the Google Sheets and reviews the DMs. That’s you or a sales manager.
• Waiting leads. Every lead who messages outside a shift waits for the next one.

With us it’s one fixed cost that stays predictable as you grow, and we go through the number on the call after we’ve looked at your DM setting. You don’t hire an AI expert or a sales manager either. My team of 10 salespeople and AI developers runs it. If you’re weighing that against an AI tool you’d run yourself, I compare the two in AI setter software vs done for you.

What happens to your setters when the AI takes over?

Nobody has to go on day one. The AI can start on a slice of your DMs while your setters keep working, and you decide how fast it takes over.

A staged start looks like this: 5% of conversations, then 10%, then 20%. Or the AI takes only new conversations, only leads from ads, or only the hours when no setter is on. Before it touches your account, you test it on a separate Instagram account. Message it as a lead, try to break it, and tell us what you don’t like. Once it’s live, you can switch it off for any single conversation with one click.

Then your setters move to the work the AI can’t do:

• Your best setter takes the follow-ups after Meta’s window, the story views, the story polls and the old hot leads. On my call with the men’s fitness coach and his two VAs, I’d already written down his best setter’s new responsibilities.
• Setters who are good on the phone move to the phones, the way Prime Train’s did.
• A setter team on another funnel can stay exactly as it is. Jordan Pham’s ads funnel kept its setter team. Only the one setter on his DMs was replaced.

The rest of a text-only setting team doesn’t carry over. Kneehab Academy had four to five DM setters. Today VAs oversee its inbox, and no human setter works its DMs.

How long does it take to switch from setters to AI?

About two to three weeks to go live, and 60 days until it’s fully dialed in. Here’s what happens in that time:

1. Week 1. An onboarding call with Emil, my co-founder, who first learns the job the way your best setter would do it. My team rescans your past DMs: where leads drop off, how long they wait, and which objections and questions come up most. You answer the common questions once, and the AI reuses your answers.
2. Weeks 1 to 2. The first version runs on a separate demo Instagram account. You message it as a lead and break it. We fix what you don’t like the same or next day, usually over two or three rounds.
3. Weeks 2 to 3. It goes live when you say so, all at once or staged.
4. Weeks 4 to 8. My team is in your DMs every day and usually spots weak conversations before you do. We start A/B tests on openers and scripts, build reminders on WhatsApp, SMS or iMessage (the target is 70 to 80% of booked calls showing up), and run a closed loop with your closers. Their verdict on every booked call keeps tightening the qualification, because the goal is booked calls with people who pay.
5. After day 60. A monthly report on every conversation, with the drop-offs, objections and most-asked questions you can turn into content. New tests start only after you approve them. There’s no minimum commitment from here.

What’s left for you: content, your calls, and feedback in a shared WhatsApp or Slack group.

What happened when Prime Train took its setters off the DMs?

Prime Train is the clearest example of the split: AI on the texting, people on the phones. Tom Baulch’s online fitness coaching business has more than 500,000 followers across Instagram and TikTok. A single lead magnet pulls 2,000 to 3,000 DM replies, and some posts get more than 3,000 comments in one day.

His setters couldn’t answer all of it. They were overworked, leads were left on read, and the lead magnets brought in close to nothing, because nobody followed up with the people who asked for them.

“It’s impossible to sift through three thousand leads.”

Tom Baulch, Prime Train, on his DMs before the AI.

He knew what it was costing him: “I was leaving a lot of money on the table.”

Tom signed with us in February 2026. My team put the AI on Prime Train’s Instagram DMs, and his setters moved off the DMs and onto the phones, where a human voice still wins. The money he saved on setter commissions went into fulfillment and ads.

About six months later, in Tom’s words: “we went from two fifty K to three fifty K a month.” That’s $250K to $350K a month at Prime Train, with the AI on the texting and his setters on the phones.

Three other clients made the same decision from different starting points:

• Kneehab Academy went from $45K to $140K a month after the AI took over all of its DM setting.
• Sam Johnston had one setter who could cover about two setting windows a day. He went from $27K to $120K+ a month in 90 days and runs zero setters and two closers today.
• Jordan Pham had one setter working his DMs about three hours a day. His DM funnel went from $0 to $117,000 in 45 days.

Each one tells it in a full interview on our client stories page.