The 6 Mistakes That Make Your AI DM Setter Useless

The most common AI DM setter mistakes are sending messages in huge robotic blocks, following a script rigidly, swaying off-script and hallucinating, and overbooking unqualified leads, and any one of them screams to your leads that they're talking to a bot. The worst thing you can have in your DMs is an AI chatbot that gives itself away like that. In the video above, I show the contrast between a cheap AI chatbot with conversion rates worse than a $3 an hour VA and an AI DM setter that outperforms a team of five intermediate setters. I'll cover what a terrible AI chatbot or DM setting team looks like, the signs that you have one, the negative impact it has on your revenue, and the best DM setting department structure you can build in 2025 and moving into 2026.

The 6 AI DM setter mistakes that make it useless

Almost every useless AI setter is making the same six mistakes:

1. It sends messages in huge robotic blocks
2. It follows the script rigidly and re-asks questions the lead already answered
3. It sways off-script and hallucinates
4. It overbooks unqualified leads
5. Its performance is inconsistent
6. It won't hand the conversation off to a human

Here is what each one costs you, and the one AI plus one human setup that fixes all six.

The results after replacing setters and chatbots

First, here are the results that show this thing works, and you can judge it yourself.

• Jaden has an Instagram page with 300k followers. He went from paying a team of five setters around $8,000 a month in commission and base salaries, with capacity for only 200 leads a day at a 15-minute speed to lead, to paying less than $4,000 a month for the entire DM setting department. That consists of one setter and one AI, with capacity for up to 1,200 leads per day, one minute speed to lead, 24/7 coverage, the same conversion rates he had before, and less management headache. All of this was done in under 30 days.
• Josh went from an average 6% booking rate with a team of five VAs to an 8 to 10% booking rate with the same framework, just by changing his team structure from five VAs to one AI and one experienced human setter. Done in under 30 days as well.
• Jera is almost the same case. He went from paying $10,000 a month for a team of five VAs who did robotic work at a 5% booking rate to an 8% booking rate now with the exact same system, all in under 30 days.

There are other successful clients I won't cover in depth here, but you get the point. This thing works, and you can see more AI DM setter results broken down call by call.

What a terrible AI chatbot or DM setting team looks like

Terrible has a lot of different meanings, so here's what it means to have one. The first example is an automated message system where you send a DM to your leads who opt in. They tell you to connect with them, or they ask a question, and your automation just replies with another stupid message. The messages get sent in one block, like a bot, the performance is inconsistent, and your DMs end up looking robotic.

The signs you have a bad AI setter

The messages are sent in huge blocks

First of all, the messages it sends are taller and wider than they have any reason to be. Humans these days can detect whether they're speaking to a chatbot, and whether you have a chatbot or not, sending those messages in large blocks makes it look very robotic. I got on a call with a coach who set up a low ticket AI chatbot in her business, and she said, "I'm getting a lot of people asking if it's AI, which I think is a bad sign." A lot of people were asking her if this is AI they're speaking with, because of those large text blocks.

It follows the script rigidly

The second thing that drops your conversions a lot is if your setters or the AI follow a script rigidly. Here's an example. Say you're a fitness coach with a script like this:

• Question one: what's your fitness goal?
• Question two: what's the obstacle holding you back from reaching that goal?
• Question three: are you open to advice? (asking their permission)
• Question four: how about we get on a call where I can help you?

The lead's response to question one might be, "My goal is to lose 5 kilograms but my lack of discipline isn't helping me. Can you help me?" They just answered three questions in one single response. They revealed their obstacle, and by asking "can you help me," they gave us the authority and permission to help them, so we can skip "hey, are you open to advice?" In that case a good AI setter pivots straight to question four and proposes the call, so it doesn't sound repetitive. A bad AI setter follows the script rigidly and keeps asking the next questions like a bot. That makes the lead feel unheard, and that's the worst thing you can have in your DMs. As another coach told me on a call, "It's not necessarily reading the conversation 100%, it does maybe 70% of the job. And if you miss something subtle, it kind of messes up the whole connection with the person." That hurts for the people on the other side of the screen.

It sways off the script

The third sign is it sways off the script a lot. Take the same script, but this time the lead says, "My goal is to lose 5 kilograms, but my lack of discipline isn't helping me. What's the method you use?" A good AI setter answers the question and continues qualifying the lead. A bad AI setter opens up more doors that lose the structure of the conversation. It says something like, "Yeah, I used the 4-day fat loss method, let me know if there are other things you want to know more about." Then the lead asks another question, the AI hallucinates another response, ends with "hope this helps," and the cycle goes on and on with no clear path. The AI is just creating different scripts in its mind and going off the main framework. As the coach put it, "It's just not acting the way it should. It's asking questions which aren't on it, there's no objective to ask that question. I thought it was going to be the answer to everything." Of course, you cannot pay a few dollars or $100 a month and expect the best AI chatbot in the world that qualifies your leads with the best conversions.

It overbooks unqualified leads

Then there's the case where it overbooks unqualified leads. As that coach said, "It's not qualifying the people correctly." That leads to terrible show rates, because it booked the wrong leads, and because the conversations the AI had weren't productive, didn't build enough trust, and didn't get the leads to show up. Her words: "The show rates have been horrific."

Its performance is inconsistent

It has inconsistent performance because it's not based on a solid framework, it sways off the script, and it books random, unqualified leads. There is no floor under its conversion rate, so your booking numbers swing from week to week and you can never trust a forecast.

It won't hand the conversation off to a human

It doesn't stop and hand the conversations over to a human. If the lead gives an objection, the AI handles it worse than my 10-year-old relative who plays Fortnite. The lead tells the AI to contact them later, and the AI keeps bouncing messages back every time the lead replies, until the lead just stops sending messages. It's a weird dynamic. And if the lead asks whether this is a scam, the AI in a robotic fashion tells the lead it's not AI. It's very ironic. All of these happen when you have a low quality AI chatbot, or a bad DM setting team, a bunch of VAs sending messages without reading context, without giving their best to qualify your leads and connect with them on an emotional level.

The negative impact on your business

The main metrics you should be tracking in your DM setting funnel are leads to call proposed, call proposed to calendar link sent, calendar link sent to call booked, and the percentage of qualified booked calls. The negative impacts that come from a bad setup are:

• Increased funnel drop-off rates due to poor AI interactions.
• Lower conversion rates, as the leads don't feel heard or understood.
• An extended conversion process that reduces overall booking percentages. The more questions you ask and repeat, the lower your conversions.
• A high volume of unqualified booked calls.
• Poor show-up rates among qualified leads, because the AI didn't have a good conversation with them.
• Damaged brand reputation from obvious automation.
• Wasted time following up with poor-fit prospects.

Here's what that translated to for that one coach. She had 23 booked calls and only six were qualified, which means the AI wasn't qualifying leads properly. That's two qualified from 271 conversations, a roughly 2% qualified booking rate, well below industry standards. It should be a minimum of 8%, and the math behind a healthy booking rate shows how far below the line that really is. Her show rates dropped from 90% with her human setters to 33% with this poorly built chatbot. There was no accurate data tracking, and a big missed revenue opportunity. We analyzed it, and her DMs could easily be generating $73.5K USD per month in revenue. Instead she was only generating a fraction of that. That's how big the impact is when you don't have the right DM setting department structure.

What is the best DM setting department structure in 2025?

First, understand the industry standard. The minimum booking rate you should get is at least 8%, and that should mostly be qualified booked calls. In our case, out of 1,000 leads we get at least 80 qualified booked calls consistently. Consistently is a big word, because if your data goes up and down, up and down, your business and your stress will reflect that. Here's how to build a high converting and everlasting DM setting system with one AI and one human setter, a split I cover in depth in ChatGPT vs human appointment setters. This system keeps getting better every day as the AI models progress, and your human setters get an amazing lifestyle because they don't have to do the repetitive work the AI does for them.

Find a winning DM setting funnel

This is the most leveraged activity you can do, because simple increases in conversion when you change your DM setting funnel structure can lead to big results if you're generating tens or hundreds of thousands of dollars from your DMs. The way to do that is by split testing different ones, or by watching my videos where I share the best DM setting frameworks you can implement. If you want a starting point, here's the best Instagram DM script for coaches.

Build an AI DM setter

I say build an AI DM setter because that's the future. AI is getting smarter day by day, it doesn't have a finite amount of energy, and it can handle 1,000, 2,000, or 10,000 leads a day. The more you improve it, the more you can scale your business without the human problems that come with it. Even the best human setter can only take a certain amount of leads, and the more leads you give them, the lower the quality of their conversations. So future-proof yourself with an AI DM setter. Build it so that:

• It can take 100 to 2,000 leads a day depending on your volume.
• It's trained on human setters' data, so it speaks and thinks like an intermediate human setter.
• Instead of asking for contact details, which causes drop-offs, it asks curiosity questions.
• Instead of repeating the same words, it acknowledges genuinely.
• Instead of speaking like a bot, it speaks like a friend.
• Instead of sending one block of text, it splits the messages.
• It sends voice messages and video messages in your own voice, which grows trust with the leads.
• It sends personalized follow-ups based on each step of the appointment setting framework and adapts to each unique lead.
• It disqualifies non-ideal avatars immediately, so you don't waste your time on sales calls with the wrong leads.
• It handles the post-booking process and handles objections.

You also need to handle errors and the new API updates that come up every week to keep the system running smoothly, because you will get a lot of errors. This is software, that's normal. But it's worth it, because you get the advantages of scalability, instant responses, and AI models that keep getting smarter. I break the build down further in AI DM automation for coaches.

Hire and train a human DM setter

After you have a winning DM setting funnel and a great AI DM setter, hire and train a human setter so they can take over the conversations that drop off from AI. Onboard them with all the info they need, then hop on daily calls for two weeks straight to review performance and feedback. After that, quality control the AI and the setters' performance daily for another two weeks. They send screenshots in the Slack channels showing where the AI can improve, and you send them feedback on what they can improve too. Then build the right data tracking systems to know the performance of the AI and the human setters distinctly, and hire a developer to help you build a better, more efficient version of the AI setter, because it can always be better.

These are the steps we followed to achieve an average of 12% booking rates across clients, consistent performance (a big, big thing), one minute speed to lead, 24/7 coverage, 2k lead capacity per account, and 80 to 90% automation with AI.

What your next steps are

If you've realized already, none of these steps are easy to do. That's why it has taken us more than two and a half years to continuously develop and refine all the systems we have, including the sophisticated ManyChat sequences, hundreds of automations, highly converting and trust-building frameworks for all niches, a custom-built AI setter, a detailed data tracking system with the CRM, KPI trackers, and effective objection handling databases. To be honest, we love building out amazing appointment setting systems. My personal vision is to master the skill set so I can deliver the best appointment setting systems in the industry. Following these steps might look simple here, but there are lots of nuances in this thing. If you want a fuller picture of the model, here's how AI appointment setting for coaches actually works.

As a last message from me to you, don't let your DM setting department ruin your good moments. Cheers.