setscale

AI appointment setter: the 3 kinds, what each one costs, and the calls it can and cannot book

September 27, 2026·19 min read

An AI appointment setter is software that answers the people who message a business, asks the qualifying questions, and books the ones who fit into a calendar. Three different products are sold under that name, each is metered differently, and the rules of the channel decide which calls the setter is allowed to book at all. This guide sorts the three kinds, prices the one we can price from a page we can cite, and lists the calls a DM setter can and cannot book under Meta's own documentation, read on 27 September 2026.

TL;DR

  • Three kinds share the name: the DM setter answers Instagram, WhatsApp and Messenger conversations; the voice setter phones leads; the outbound setter runs email or SMS sequences. Each is metered on something different, and the meter matters more than the headline price.
  • On Instagram and Messenger a setter cannot start a conversation: Meta's documentation states that the conversation must be initiated by the person, then gives the business 24 hours to reply. Outside that window, the only tag that fits an escalation is the human agent tag, defined for a business representative replying manually within 7 days.
  • On WhatsApp the setter can reach out first, with a pre-approved template charged per delivered message. Replies inside the 24-hour customer service window are free until 30 September 2026; from 1 October 2026 Meta charges service messages per message.
  • SetScale meters the DM kind per AI message: 19, 97, 297 and 797 US dollars a month for 300, 1,500, 5,000 and 15,000 messages, as displayed on the pricing page on 27 September 2026. One AI reply is one message.
  • The number to decide on is the cost per call that holds, not the monthly fee. The cost per booked call calculator computes it from your own volumes and your own show-up rate.

Table of contents

The three kinds of AI appointment setter

The name covers three products that share one promise, a calendar filled without a human reading every message, and almost nothing else. They differ in where the conversation happens, in who is allowed to start it, and in what the vendor counts when it bills you.

Kind Where it works Who starts the conversation What it books What to read on the vendor's pricing page
DM setter Instagram DMs, WhatsApp, Messenger, replies to comments The lead, on Instagram and Messenger; the business too on WhatsApp, with a template A slot proposed inside the conversation The unit: messages, conversations or contacts, and what happens above the quota
Voice setter Phone calls, inbound or outbound Usually the business A slot agreed during a call Whether minutes are billed on top of the monthly fee, and what a qualifying call lasts
Outbound setter Email and SMS sequences The business A reply that turns into a booking link Whether the price is per contact, per send or per seat, and who supplies the list

Two other names appear in the same searches. "AI setter" is the first kind, shortened. "DM setter" names the channel rather than the technology: a setter, human or automated, who works in direct messages. The rest of this guide is about that first kind, because it is the one we build and the one whose rules and prices we can document from primary sources.

The rules a DM setter works under

Every DM setter, whatever the vendor claims, operates inside the rules Meta publishes for its messaging APIs. The four that shape what a setter can do were read on Meta's developer documentation on 27 September 2026.

The person opens the door. Meta's guide to sending messages on the Messenger Platform states: "To send messages to a person on Messenger or Instagram, the conversation must be initiated by that person" (Send a message, Messenger Platform, page updated 11 August 2026). The actions that open a conversation are listed on the same page: a message, a click on a call-to-action button, a click on a Click-to-Messenger ad followed by a message, an ig.me link, a reaction to a message, a comment on a post.

The door stays open 24 hours. The same page defines the standard messaging window as "the 24 hour time period in which you are allowed to send a message to a person", and adds that messages inside it may contain promotional content. The Instagram Messaging overview (updated 26 June 2026) says the same from the Instagram side: your messaging app has 24 hours to respond to a message, and "if a human agent needs more time to respond, you can add a tag to your response".

Outside the window, tags, and the tags are not for selling. A tagged message "must include a message tag that matches the allowed use case for the tag and contains non-promotional content". The human agent tag "allows a business representative to manually respond to a person's messages within a 7-day period". Message tags "may not be used to send promotional content, including but not limited to: deals, offers, coupons, and discounts", and misuse "may result in restrictions on the Page or Instagram account's ability to send messages". Sponsored Messages and One-Time Notifications, the two Messenger tools for reaching people after the window, "are not available for Instagram Messaging API".

The setter must say what it is, and hand over to a human. Apps using the Instagram Messaging APIs "must have an escalation path to a human agent" (overview, 26 June 2026). When required by law, an automated experience must disclose that the person is talking to an automated service at the beginning of a conversation, after a significant lapse of time, or when the chat moves from a human to an automation; Meta names California and Germany as markets where this deserves particular attention, and recommends the disclosure everywhere as best practice.

WhatsApp follows a different set of rules, with the same 24-hour logic. On the WhatsApp Business Platform, template messages "are the only message type that can be sent outside of a customer service window, to reach out to users", and a customer service window "opens and resets with each user message" (pricing page, updated 1 July 2026, and its upcoming pricing updates page, updated 25 August 2026). The number setup, the shared messaging limits and the opt-in requirement are covered in our eight rules for the WhatsApp Business API, and the Instagram side of the same question, the per-account rate limits, in Instagram messaging API rate limits per client account.

The calls a DM setter can book, and the ones it cannot

Put the rules above next to a calendar and the list writes itself.

Calls it can book

  1. The person who messages first. On all three channels, the message opens the window, the setter qualifies inside it, and every new message from the person opens the window again. This is the core of the job and the only case where the setter is fully autonomous.
  2. The person who comments on a post, reel or ad. Meta's Private Replies documentation (updated 2 July 2026) allows a single private message to a commenter within 7 days of the comment; the message lands in the Inbox if the person follows the account and in the Requests folder if not. The call gets booked only if the person answers, because "only when the Instagram user responds to the private message can you continue the conversation within the 24-hour messaging window".
  3. The person who clicks a Click to WhatsApp ad. A reply within 24 hours is free and opens a 72-hour free entry point window during which "you can send any type of message to the user at no charge" (WhatsApp pricing page, 1 July 2026). Our piece on click to WhatsApp ads covers what has to be in place before the first click.
  4. The person on your WhatsApp list who opted in. The setter opens with a marketing template, charged per delivered message at the rate of the recipient's country, and continues free-form once the person replies.
  5. The reminder before the call, on WhatsApp. A reminder sent outside the window is a template message and is charged per delivery; a utility template inside an open window is free until 30 September 2026 and charged from 1 October 2026.

Calls it cannot book

  1. Anyone on a cold Instagram or Messenger list. There is no outbound. The conversation must be initiated by the person, and the API offers no template, tag or paid message on Instagram that changes that.
  2. The commenter who never answers the private reply. One message is the limit, and the thread only continues if the person writes back.
  3. The Instagram lead who went quiet more than 24 hours ago. A promotional follow-up is not allowed under any tag. The human agent tag exists for a business representative answering manually within 7 days, which makes that follow-up a human's job, not an automation's.
  4. The WhatsApp lead outside the window, with a free-form message. Only a template can be delivered, and it is charged.
  5. The person who never gets told they are talking to software, in a market where the law requires it. The disclosure is a condition of operating, not a courtesy.

A vendor who promises calls from the second column is either describing a human on the late shift or describing a practice that puts the account at risk. Ask which.

What each kind costs: read the meter before the price

A monthly fee tells you very little until you know what is counted. Three meters exist, and one of them comes with platform charges from Meta that no vendor controls.

The DM setter is metered in messages, and on WhatsApp Meta bills its own line. On Instagram and Messenger, the platform itself is free: "The Messenger Platform is free to use" (Messenger Platform overview, updated 23 March 2026). On WhatsApp, Meta bills per delivered message since 1 July 2025: marketing templates always, utility templates outside the window, and from 1 October 2026 every service message a business sends inside the window, at the same rate as the utility and authentication messages of the market, with no volume tiers (upcoming pricing updates page, 25 August 2026). That page gives one rate as an example: 0.68 US cents per delivered message to users in Brazil, the utility and authentication rate as of 1 July 2026. It also notes that a business using a third-party AI solution "typically incur separate charges for AI usage from the AI provider and message delivery from the messaging solution provider". The full cost of a WhatsApp reply is therefore the vendor's message plus Meta's delivery.

Meta's own AI agent gives the clearest benchmark for what one automated reply costs. Since 1 August 2026, Meta Business Agent messages on WhatsApp are charged at "one global rate of $2.00 USD per 1 million (M) tokens", and Meta states that one message typically consumes 20,000 to 25,000 tokens, approximately 4 to 5 US cents per message; a simple enquiry of 4 messages is estimated at 16 to 20 cents, a complex interaction of 10 messages at 40 to 50 cents (same page). That is Meta pricing Meta, delivery included, and it is the number to hold any DM setter against.

The voice setter's cost is driven by talk time. Before comparing it to anything, read whether minutes are billed on top of the monthly fee, whether unanswered calls count, and what a qualifying call lasts on average. A per-minute rate multiplied by an average call length, plus the fee, gives the monthly total; nothing else on the page does.

The outbound setter's cost is driven by the list it works. Read whether the price is per contact, per send or per seat, and whether the list is yours or theirs. On WhatsApp, an outbound sequence is a marketing template per delivered message, which is Meta's line again, before the vendor's.

We quote no other vendor's price in this guide, on purpose. Prices change on a quarterly calendar on Meta's side and without one on most others, and a figure copied from a competitor's page in September is a wrong figure by December. The one grid we can vouch for is ours, dated below.

SetScale pricing as displayed on 27 September 2026

The prices below are those shown in the pricing section of the SetScale home page (setscale.io/#pricing) on 27 September 2026. The page is the reference: if it and this article disagree, the page wins.

Plan Monthly price AI messages included Users Client accounts What changes at this tier
Starter 19 USD 300 1 1 Fast engine, a lighter model on every reply; 3 connected channels
Pro 97 USD 1,500 Up to 3, with roles 1 Max engine; roles and permissions; one calendar per closer with round-robin; per-closer reporting
Scale 297 USD 5,000 Unlimited Unlimited Per-client reporting, branded reports, read-only client access, white-label, API and webhooks, unlimited channels
Business 797 USD 15,000 Unlimited Unlimited Dedicated onboarding

The meter is stated on the page as "One AI reply = one message". Dividing the monthly price by the messages included gives the price of one AI reply at full use of the quota: 19 divided by 300 is 6.3 cents on Starter, 97 divided by 1,500 is 6.5 cents on Pro, 297 divided by 5,000 is 5.9 cents on Scale, 797 divided by 15,000 is 5.3 cents on Business. On WhatsApp, Meta's delivery charge comes on top from 1 October 2026; on Instagram and Messenger, nothing does.

Every plan runs the same setter: qualification on your questions, booking on your available slots, automatic follow-ups by text or voice note, reminders 24 hours and 2 hours before the call, incoming voice notes handled, human takeover in the inbox, AI disclosure and STOP opt-out, a test simulator before going live, keyword triggers in DMs, comment replies with a public reply plus a DM, story replies, and WhatsApp campaigns on approved templates. What the tiers change is volume, the AI engine, the number of people who log in, and the number of separate businesses the account can run.

Cost per held call: run your own numbers

The monthly fee is not the number to buy on. The number is what one call that actually holds costs you, and the gap between calls booked and calls held is your show-up rate. The cost per booked call calculator does this arithmetic from your own figures: DMs received per month, the share that becomes a qualified conversation, the share of qualified people who book, the share who turn up, the minutes you spend per DM today and what your hour is worth. It returns the calls held per month, the SetScale cost per held call on the plan you pick, the hours you currently spend in your DMs and what those hours are worth.

Two design choices matter for a fair comparison. The denominator is the call that held, never the call that was booked, because a calendar full of no-shows costs the same time as a full one and returns nothing. And nothing in it is a market average: every input is yours, and the only values that come from us are our prices and quotas. Use the same denominator for a voice or outbound setter. Take the vendor's monthly total at your volume, divide it by the calls that held, and the three kinds become comparable on the one number that pays.

One practical check while you are there: a conversation uses several AI messages, so confirm that your monthly DM volume fits inside the 1,500 messages of Pro before you assume the Pro price. If it does not, the Scale tier is the honest comparison.

Four layers to check before you buy

The kind and the price settle the first decision. The second is whether the tool holds up when there is more than one person, one calendar or one business behind it. Four layers decide that, and each has a line in the grid above.

  1. Isolation. Each business you run needs its own container: script, conversation history, calendar links, reporting. Onboarding a client should mean creating a workspace, not creating a new subscription. On the grid: one client account on Starter and Pro, unlimited on Scale and Business.
  2. People. The unit of team use is the seat, and seats need different rights: an admin who configures, a manager who reads, a closer who takes over, a client who only views. On the grid: one user on Starter, three with roles on Pro, unlimited from Scale, with read-only client access from Scale.
  3. Flow. A qualified lead has to land as a slot in the right calendar, and a human who steps in has to receive the conversation, not a summary. On the grid: human takeover on every plan, one calendar per closer with round-robin from Pro. The pay and handoff side of that flow is in our setter closer model, and the questions that decide who is qualified in the lead qualification framework for DM teams.
  4. Proof. Numbers per closer, per client and per period, exportable. On the grid: dashboard and CSV export on every plan, per-closer reporting from Pro, per-client and branded reports from Scale.

If you are a coach with one account, layers one and four are what you use; two and three become real the day you hire a closer. If you run several businesses' DMs, all four are the product.

Where a human setter still belongs

Meta's rules make the case better than any vendor: an app on the Instagram Messaging APIs must offer an escalation path to a human, and the only tag that survives the 24-hour window is the one built for a business representative answering by hand. A setter that qualifies at inbox speed and a human who takes the judgement calls, the objections and the 7-day follow-ups is the configuration the documentation is written for. If you are hiring or training that human, our 30-day ramp for appointment setters is the plan, and the comparison between an agency, an in-house team and software is in what appointment setting services cost.

FAQ

Can an AI appointment setter message people first on Instagram? No. Meta's documentation states that on Messenger and Instagram the conversation must be initiated by the person. On WhatsApp a business can open with a pre-approved template, charged per delivered message.

Does the setter have to say it is automated? When the law of the person's market requires it, yes: at the start of the conversation, after a significant lapse of time, or when a human hands over to the automation, with California and Germany named by Meta. Meta recommends the disclosure everywhere. SetScale's grid lists AI disclosure and a STOP opt-out on every plan.

What happens to a lead who stops answering for two days? On Instagram, no promotional message can be sent under any tag; the human agent tag lets a business representative reply manually within 7 days. On WhatsApp, a template can be sent, and it is charged.

What does a WhatsApp reply cost from 1 October 2026? A per-message charge at the utility and authentication rate of the recipient's market, with no volume tiers, on top of whatever the AI vendor charges. Until 30 September 2026 those replies are free inside the 24-hour window. The detail is in what a WhatsApp conversation costs.

How many AI messages does one held call take? It depends on your script and your audience, and no market average would be honest. Count the messages in ten of your own qualified conversations, multiply by your monthly volume, and check the result against the quota of the plan.

Is an AI setter the same thing as a DM setter? An AI setter is the category, whatever the channel. A DM setter is a setter working in direct messages, human or automated. The first kind in this guide is both at once.

Conclusion

Three products, one name. The DM setter works inside a 24-hour window the lead opens, on channels where Meta publishes the rules and, on WhatsApp, the prices. The voice and outbound kinds count minutes and contacts instead, and the only way to compare them is the cost of a call that held. Read the meter, run your own numbers, then check the four layers before the price.

If you want the setter side to run at inbox speed, SetScale runs it for you: it replies, qualifies and books on your calendar, on Instagram, WhatsApp and Messenger. You start solo, on one account; when you grow, you add closers, accounts or client spaces without changing tools. Pro and Scale come with a 7-day free trial, card required, nothing charged before day 8. Start your 7-day free trial