On this page
AI appointment setting cost depends on the pricing model: a software licence, an agency retainer, usage-based fees or pay per booked appointment. The fair comparison is a UK in-house setter, who costs roughly £35,000-£45,000 a year once salary, employment costs and tools are added. Setterly charges per booked, qualified appointment, priced on your average job value.
That gives you the shape of the answer. The more useful point is that the risk of overspending is almost entirely a function of how you pay. Choose the wrong pricing model and you carry all the risk. Choose the right one and the vendor carries it.
This piece walks through the four pricing models, what hiring a setter really costs, why unclear ROI kills AI budgets, and a simple framework for budgeting that cannot be burned on a pilot.
The four pricing models for AI appointment setting
Every AI product you will be quoted charges in one of four ways. Each one distributes risk differently between you and the vendor.
| Pricing model | What you pay for | Who carries the risk |
|---|---|---|
| Licence or subscription | Access to software, per seat or per company | You: same fee whether it books anything or not |
| Retainer or managed service | Activity: setup, management, reporting | You: the invoice arrives either way |
| Usage-based | Messages, conversations or minutes | You: a struggling system can run up a large bill |
| Pay per booked appointment | Outcomes: qualified appointments in your diary | The vendor: no appointments, no cost |
1. Licences and subscriptions
A fixed monthly or annual fee per seat or per company. Predictable, easy to budget, and completely disconnected from results. If the tool sits unused or underperforms, you pay the same, and you still supply the staff time to run it. Fine for commodity tools. Dangerous for big platform commitments where the ROI case is unproven.
2. Retainers and managed services
An agency or consultancy runs the AI for you for a monthly fee. You are paying for activity. Results may follow, but the invoice arrives either way. Retainers concentrate the attribution problem, because the monthly report has to justify the monthly fee, which encourages activity metrics over revenue metrics.
3. Usage-based pricing
Pay per message, per conversation or per minute. Common with AI infrastructure. The trap is that usage scales with activity rather than outcomes, so a poorly performing system can generate a large bill precisely because it is thrashing away without converting. Budgeting is also hard, because you cannot forecast usage until you are already committed.
4. Pay per booked appointment
You pay only when the AI produces the outcome you actually wanted: a booked, qualified appointment. The vendor funds the technology, the setup and the running costs, and earns only from output. This is the only model where the ROI question is answered by the invoice itself.
Setterly works this way. You pay per booked, qualified appointment, priced on your average job value, with no setup fee and no retainer. Optional CRM integration is quoted separately. The detail is on our pricing page, and the logic is the same one we set out in pay per appointment vs pay per lead.
What hiring a setter really costs
A human appointment setter or SDR in the UK typically costs £25,000 to £35,000 in base salary before commission. Add employer's NI, pension, a phone system, a CRM seat, and recruitment and management time, and one setter is a £35,000 to £45,000 annual commitment.
Then add what nobody budgets for: a ramp-up period, holidays, sickness, turnover and the chance the hire is simply not very good at it. One person works around 1,700 hours a year, one conversation at a time, and is off when 56-60% of dealership leads arrive in the evenings and at weekends.
All of that is paid before appointment one. Compare it with a model where the worst case is a bill of zero. Our guide to AI appointment setting covers where AI beats hiring and where humans still win.
Why unclear ROI kills AI budgets
Picture the meeting that kills an AI project. It is twelve months in. Say the retainer has cost £3,000 a month, £36,000 in total. The dashboard shows conversations handled, response times improved, engagement up. Then someone asks what revenue it produced, and the honest answer is a shrug dressed up as a methodology discussion.
That meeting is why unclear ROI attribution sits alongside weak governance and missing observability as a leading driver of the 40%+ cancellation risk for agentic AI projects. The projects did not necessarily fail technically. Nobody could connect the spend to money.
AI sales projects can be brutally measurable when structured properly. The winners pick projects where every outcome traces to a named lead, a booked appointment and a closed deal. Attribution is designed in before a penny moves.
How pay per appointment removes the budget risk
Here is what outcome pricing does to your budget mechanics.
Your worst case becomes zero. If the AI books nothing, you pay nothing. There is no scenario where you write off a pilot budget, because there was no pilot budget.
Attribution is automatic. Every pound you spend maps to a booked appointment with a named lead from your own CRM. When the finance director asks what the AI produced, the answer is a list of appointments and the revenue they closed into.
Cost scales with revenue, in the right direction. A licence costs the same in a bad month as a good one. A pay-per-appointment bill is only ever large in months where the system filled your calendar.
Real numbers show the shape. A single-branch UK window and door installer booked 22 appointments in its first 30 days, closed £31,483 of work and paid £1,520 in appointment fees. Cloudy2Clear, a UK misted double glazing repair franchise, re-engaged 6,769 old enquiries and booked 673 appointments, generating £639,350 in pipeline. CarFinance247, a UK car finance broker, booked 114 appointments over Christmas from leads its sales team had already worked. Across 30+ businesses the model has produced 3,000+ appointments. The detail is in the window and door case study and on our results page.
A simple framework for budgeting AI appointment setting
Four steps, in order.
Step 1: Put your first AI spend on a results-priced project
Start with one project priced on outcomes. For lead-generating businesses, the obvious candidate is database reactivation: an AI salesperson working the old leads already in your CRM, paid per appointment it books. Budget required to start: nothing. Downside: none, the leads were dead anyway. Our database reactivation guide covers the method.
Step 2: Measure appointments and revenue, nothing softer
Track three numbers only: appointments booked, show rate and revenue closed. If a metric cannot be converted to pounds, it does not belong in the budget review. This is how you stay out of the attribution trap.
Step 3: Set an expansion trigger, in writing
Decide now what result unlocks further AI spend. For example: "If reactivation returns over 3x within two quarters, we extend AI to live lead follow-up." This turns your AI budget from a hopeful guess into a conditional plan. The one-page version is in AI for small businesses: where to start.
Step 4: Expand from proof, and stay outcome-weighted
With a hard ROI number banked, add the second and third projects: AI on fresh enquiries through speed to lead, then analytics or content tools funded by reactivation revenue. Keep the portfolio weighted towards outcome-priced spend. The mechanics of the first project are on our how it works page.
Frequently asked questions
How much does AI appointment setting cost?
It depends on the pricing model. Licences and retainers charge a fixed monthly fee, usage-based tools charge per message or conversation, and pay-per-appointment providers charge only for booked appointments. Setterly charges per booked, qualified appointment, priced on your average job value, with no setup fee or retainer.
Is AI appointment setting cheaper than hiring a setter?
Usually, and it carries less risk. A UK setter costs roughly £35,000 to £45,000 a year all in, paid regardless of results, and works one conversation at a time during office hours. AI priced per booked appointment costs nothing if it books nothing and covers every hour of the week.
Why do AI pilots waste so much budget?
Pilots are usually structured as fixed spend against vague success criteria, which is why unclear ROI attribution is a leading cause of the 40%+ cancellation risk for agentic AI projects. Outcome pricing removes the problem: with no fixed spend, there is nothing to waste.
Which AI pricing model is safest for a small business?
Pay on results, because the vendor carries the performance risk. Licences, retainers and usage pricing all charge you whether or not revenue follows. Paying per booked appointment means every invoice is, by definition, already justified.
Works for: Windows and doors, Solar and EV charging, Car dealerships, Car leasing
