TL;DR: Vapi, GHL, and Cal.com stack into a lead-qualification and booking system that runs under $500 a month. That beats a fully loaded SDR by a wide margin: the Bridge Group's 2025 SDR benchmark report puts median SDR on-target earnings at $80,000 a year, and fully loaded in-house cost runs $104,000 to $205,000 a year once you add payroll tax, benefits, tooling, and ramp time. This build answers an inbound lead in under a minute, asks three qualifying questions, and books a confirmed appointment on your calendar. No SDR required.
Key Takeaways
- Vapi answers the phone, GHL runs the CRM and follow-up, Cal.com owns the calendar. Three tools, one job: turn a ringing phone into a booked appointment.
- Realistic all-in cost for agency-level call volume runs $200 to $450 a month: Vapi usage plus a GHL Starter or Unlimited plan plus Cal.com's free or Teams tier.
- A fully loaded in-house SDR costs $9,000 to $17,000 a month. Even an outsourced SDR runs $3,500 to $5,500 a month. This build pays for itself inside 30 days for most agencies.
- This is not a chatbot widget. It is a voice agent that answers, qualifies against your criteria, and books the meeting without a human touching the call.
The Problem: Your SDR Line Item Is a Leak
Every owner-operator agency hits the same wall. Leads come in through the website, a paid campaign, or a referral form. Someone has to answer, qualify, and book. That someone costs real capital.
The Bridge Group's 2025 report, built from 351 B2B companies, puts median SDR on-target earnings at $80,000 a year on a 68:32 base-to-variable split. Add payroll tax, health benefits, and a tooling stack and you land at $104,000 to $205,000 fully loaded, per multiple 2025-2026 benchmark studies. That is $9,000 to $17,000 a month for one seat. And that seat only hits productivity after a three-to-four-month ramp, with 40% annual attrition industry-wide.
Outsourcing does not fix the math. Division50's SDR cost calculator puts a dedicated US-market outsourced SDR at $3,500 to $5,500 a month, still north of $40,000 a year. One Texas SaaS startup profiled by Digital Patron was burning $20,000 a month on an underperforming four-person SDR team, at $1,200 per booked meeting.
In damage control, you assign a watchstander to every compartment. No compartment goes unmanned, because an unmanned compartment is how a ship sinks. Your inbound lead flow is a compartment.
If nobody answers inside five minutes, the lead floods and you lose the ship to a competitor who answered faster. An AI appointment setter is your lead-response watchstander. It never leaves its post, and it never calls in sick.
The Stack: Vapi + GHL + Cal.com
Three tools, three jobs, one budget line under $500.
Vapi is the voice layer. It answers or places the call, runs your qualifying script, and hands off a structured result. Vapi's platform fee is a published $0.05 per minute, with speech-to-text, the language model, text-to-speech, and telephony billed at cost on top.
Realistic all-in cost lands around $0.10 to $0.30 per minute depending on the voice and model you pick, according to multiple 2026 cost breakdowns. There is no required monthly subscription on Vapi's self-serve Build plan. You pay for the minutes you use, and the platform includes ten concurrent calls before you need to add capacity.
GHL (GoHighLevel) is the CRM and workflow engine. It holds the contact record, triggers the call, logs the qualifying answers, and runs the follow-up SMS and email sequence if the lead does not book on the first pass. GHL's pricing page lists three tiers: Starter at $97 a month with three sub-accounts, Unlimited at $297 a month with unlimited sub-accounts and API access, and Agency Pro at $497 a month with SaaS reseller mode. Most single-agency operators only need Starter or Unlimited.
Cal.com is the calendar. It holds your real-time availability, exposes a booking API, and confirms the appointment the second the AI closes the qualification. Cal.com's free plan includes unlimited event types, unlimited calendar connections, and full booking automation.
That tier alone covers most solo-operator builds. The Teams plan runs $12 per user per month if you need shared availability across a small booking team, with round-robin routing so a qualified lead lands on whichever closer has the next open slot.
You can also run the qualification and follow-up entirely inside GHL's own AI Employee add-on, which GHL prices at $50 to $97 a month per location. That trades some of Vapi's flexibility for one less integration to maintain. Most operators building a first version are better served keeping the three tools separate, since each one does its job better on its own.
Add it up: Vapi usage at moderate agency volume runs $75 to $300 a month, GHL Starter or Unlimited runs $97 to $297 a month, and Cal.com costs $0 on the free tier. Total: well under $500, with room to spare.
The Step-by-Step Build
Step 1: Build your Cal.com calendar and event type. Create the "Discovery Call" or "Strategy Call" event type. Set your real availability, buffer time, and time zone rules. Grab your Cal.com API key from account settings.
Step 2: Set up your GHL pipeline and workflow. Create a pipeline stage for "New Lead," "AI Qualifying," and "Booked." Build a workflow that fires the moment a form fills or a call comes in, tagging the contact and triggering the Vapi call within seconds, not minutes.
Step 3: Write your Vapi assistant script around three qualifying questions. Keep it tight. A typical agency script runs: (1) What service are you looking for? (2) What is your monthly budget range? (3) When are you hoping to get started? Three questions, thirty to sixty seconds, enough signal to route hot leads and disqualify tire-kickers.
Step 4: Connect Vapi to GHL. Use Vapi's webhook to push the call transcript, the qualifying answers, and an intent score into the GHL contact record the moment the call ends. This is your qualification log, and it is what turns a phone call into structured pipeline data.
Step 5: Connect Vapi to Cal.com for direct booking. Vapi's function-calling layer hits the Cal.com API v2 booking endpoint mid-call. When a lead qualifies, the assistant checks live availability and books the slot on the call, no back-and-forth, no "I will follow up with some times."
Step 6: Build the escalation and fallback path. Any lead that asks for a human, or any answer that trips a disqualifying rule, routes to a Slack or SMS alert to you. The AI handles the repeatable ninety percent. You handle the judgment calls.
Step 7: Test with fifty real or seeded calls before going live. Score the qualifying accuracy, the booking success rate, and the transcript quality. Fix the prompt, not the tool, when something breaks.
Cost Breakdown at Agency Volume
| Line item | Monthly cost | |---|---| | Vapi (moderate volume, 1,500–3,000 minutes) | $150–$400 | | GHL Starter or Unlimited plan | $97–$297 | | Cal.com (Free or Teams tier) | $0–$36 | | Total | $250–$450 |
Compare that to $9,000 to $17,000 a month for one in-house SDR, or $3,500 to $5,500 a month for an outsourced one. The Reprise AI case study on a twelve-agent Tampa real estate team shows the shape of the payback: an AI voice qualifier tripled lead-to-appointment conversion from 3.8% to 11.9% and drove a six-figure gross commission lift inside the first year. Payback ran under sixty days.
A nine-person B2B agency profiled in a separate case study replaced a churned SDR hire with an AI booking pipeline and tripled qualified meetings from 12 to 41 a month, with payback in 27 days. That is the number to hold onto. Not the technology. The number.
What to Watch For
Watch your concurrency limits. Vapi's Build plan includes ten concurrent calls before you pay extra per line. If your agency runs paid ad spikes, budget for overage.
Watch your compliance exposure. Outbound AI calling triggers TCPA rules in the US. Inbound answering does not carry the same restrictions, so keep this build on the inbound side until you have counsel review your outbound plan.
Watch the qualifying script drift. Review a batch of transcripts weekly for the first month. A qualifying question that sounds fine on paper can confuse callers on a live line. Fix it fast.
Watch for the founder becoming the fallback queue. If every escalation lands on your phone, you have rebuilt the SDR bottleneck with better branding. Route escalations to whoever owns client intake, not to the owner by default.
Watch your voice choice against your brand. A premium AI voice on ElevenLabs costs more per minute than a standard voice, but a robotic-sounding qualifier will cost you bookings before it ever saves you a payroll dollar. Budget the extra few cents a minute; it is cheap insurance on your close rate.
Doctrine Connection: Ownership Beats Wages
The Owner's Exit Engine framework says an agency becomes sellable the day its revenue-producing functions run without the founder standing watch. A wage-based SDR is a liability on your income statement and a dependency on your org chart. An AI appointment setter you built once is an asset on your balance sheet, and it runs the same shift whether you are in the office or on a plane.
Wages scale with headcount. Ownership scales with usage. Every month this system books appointments without a payroll line, you are converting a recurring cost into owned infrastructure.
Owned infrastructure is what a buyer pays a multiple on. A buyer does not pay a premium for your SDR's Rolodex. A buyer pays a premium for a system that keeps booking appointments after you are gone, because that system proves the agency is operator-independent.
Frequently Asked Questions
Does an AI appointment setter really replace a full SDR, or just some of the work? It replaces the repetitive, time-sensitive part of the role: answering fast, asking the same three qualifying questions every time, and booking without back-and-forth. Case studies from Deburise, Digital Patron, and Reprise AI all show the same pattern: the AI handles first response and qualification, and any remaining human capacity moves into closing, not answering the phone.
What happens if a lead asks a question the AI cannot answer? Build an escalation rule into the Vapi assistant. Anything outside the qualifying script routes to a live transfer or a flagged follow-up task in GHL. You are not removing the human from the sale. You are removing the human from the phone-answering queue.
Do I need developer skills to connect Vapi, GHL, and Cal.com? Basic webhook and API-key setup is required, and most agency owners can follow the documented steps or hand a half-day task to a freelance automation builder. None of the three platforms require custom code for a standard qualify-and-book flow.
How fast is the real payback period? Case studies in this space report payback in 27 to 60 days once the system is live and tuned. At a build cost under $500 a month against an SDR line item of $3,500 to $17,000 a month, the arithmetic works in your favor from month one.
Jeff Barnes has no personal position in any company, tool, or platform named in this article. DEMG has no current commercial relationship with any party mentioned. DEMG provides marketing strategy and AI operations guidance, not investment advice. Results described are illustrative and not guaranteed.