Should you choose AI or outsourced appointment setting?

The monthly fee does not tell you which option is better value. Compare the work included, the appointments produced and the supervision your team still provides. AI, an outsourced service or a combination can each make sense; the right choice depends on the task and the results you can verify.

In the hypothetical worksheet below, one workflow costs $2,400 a month and another costs $4,800. Yet both cost $60 per attended, qualified appointment. The first spends less; the second produces more appointments. Neither price alone tells you which suits your business.

Those are invented AUD figures, not vendor quotes or forecasts. The worksheet explains the GST, setup and lead-cost assumptions. Before using the calculation, make sure the two proposals cover the same job.

This guide focuses on qualification and booking for leads you supply. Finding new prospects is a different service and should be priced separately. Vendita’s appointment-setting workflow uses SMS, live Google or Outlook availability and direct booking after a time is agreed. If you need calls, evaluate that channel explicitly; the appointment setter guide explains the distinction.

What work is included in each option?

Write down the job you want done before comparing proposals. Then check which tasks each supplier includes and which remain with your team. An agency may supply people, automation and campaign management; a software subscription may require you to run more of the process yourself.

A useful scope might be: “Follow up our website quote enquiries, establish service fit and book an inspection into our calendar.” Give both suppliers that same brief.

Use the table below to compare the proposals. Ask about coverage rather than assuming an agency works only office hours or that software availability includes overnight human support. Record additional work, such as prospect research, separately so it does not distort the price comparison.

Scope questions to put in both proposals

Lead supplyDo we supply all contacts and eligibility records?

Outsourced service: askAre leads supplied, researched or purchased, and who verifies eligibility?

ChannelsSMS, voice or both; what is demonstrated in production?

Outsourced service: askWhich channels do staff and automation actually use?

CoverageWhat are the sending rules, support hours and outage process?

Outsourced service: askWhat hours, response commitments and backup cover are contracted?

QualificationWhich rules are enforced, and what happens with unclear answers?

Outsourced service: askHow are setters trained, assessed and updated when criteria change?

BookingIs a live event created, or is a link or request passed to someone else?

Outsourced service: askWho confirms the time and handles calendar conflicts?

ExceptionsWho receives the handoff, and how is ownership acknowledged?

Outsourced service: askWhich questions reach your team, and within what timeframe?

OperationsWho maintains business information, integrations and review queues?

Outsourced service: askWhich management, reporting and CRM tasks are included?

AreaAI workflow: askOutsourced service: ask
Lead supplyDo we supply all contacts and eligibility records?Are leads supplied, researched or purchased, and who verifies eligibility?
ChannelsSMS, voice or both; what is demonstrated in production?Which channels do staff and automation actually use?
CoverageWhat are the sending rules, support hours and outage process?What hours, response commitments and backup cover are contracted?
QualificationWhich rules are enforced, and what happens with unclear answers?How are setters trained, assessed and updated when criteria change?
BookingIs a live event created, or is a link or request passed to someone else?Who confirms the time and handles calendar conflicts?
ExceptionsWho receives the handoff, and how is ownership acknowledged?Which questions reach your team, and within what timeframe?
OperationsWho maintains business information, integrations and review queues?Which management, reporting and CRM tasks are included?

These are due-diligence questions, not claims that every AI product or agency has the same capabilities.

How should you compare the true cost of appointment setting?

Include the full cost of operating the agreed workflow: supplier fees, usage, setup, integrations and your team's time. Then divide by appointments that meet your qualification criteria and are attended. Track total outcomes too: a low cost per appointment can conceal a workflow that delivers too little useful volume.

For software, ask whether usage is counted by contacts, messages, SMS segments, call minutes or appointments. Include overages, additional numbers and any required integration work. Price the time your team spends updating information, reviewing conversations and handling escalations.

For an agency, confirm the retainer, setup fee, performance charges, minimum term and the definition of a billable appointment. Clarify how no-shows, duplicates, reschedules and appointments outside your criteria are handled. Management time may fall, but do not assume it becomes zero.

Use one currency and consistent GST treatment. Spread setup costs over an explicit planning horizon for the economic comparison, and also show the upfront cash requirement. If lead acquisition costs are identical, you can show them separately; if the suppliers provide different leads, include those costs and acknowledge that you are comparing different services.

Cost per attended, qualified appointment = total relevant operating cost ÷ attended appointments that meet the agreed criteria. When that denominator is zero, report spend and zero outcomes; the ratio is undefined. Vendita's current subscriptions are on the pricing page; confirm allowances and usage charges for your proposal rather than treating this example as a quote.

Illustrative monthly cost worksheet in AUD, excluding GST

Subscription or service retainer$900

Example outsourced service$3,000

Usage or performance fees$300

Example outsourced service$600

Setup allocation$300 ($1,800 over six months)

Example outsourced service$300 ($1,800 over six months)

Internal oversight at $60/hour$900 (15 hours)

Example outsourced service$900 (15 hours)

Total monthly operating cost$2,400

Example outsourced service$4,800

Attended, qualified appointments40

Example outsourced service80

Cost per attended, qualified appointment$60

Example outsourced service$60

Cost itemExample AI workflowExample outsourced service
Subscription or service retainer$900$3,000
Usage or performance fees$300$600
Setup allocation$300 ($1,800 over six months)$300 ($1,800 over six months)
Internal oversight at $60/hour$900 (15 hours)$900 (15 hours)
Total monthly operating cost$2,400$4,800
Attended, qualified appointments4080
Cost per attended, qualified appointment$60$60

Invented figures for calculation only, not market prices, forecasts or Vendita results. Assume equivalent 500-lead cohorts, all outcomes complete, and identical acquisition costs excluded from both columns. Equal oversight hours are an example input, not a claim about normal staffing. Setup cash is paid upfront even though it is allocated across six months here.

What does the cost worksheet actually tell you?

The example has equal unit cost but different total spend and appointment volume. The software workflow needs less spending and produces fewer attended qualified appointments under the stated assumptions. Neither column proves the better purchase. The decision also depends on sales capacity, appointment value, conversion to sales and confidence in those assumptions.

Sensitivity matters. Holding the example costs fixed, 30 attended qualified appointments would put the AI workflow at $80 each; 60 would put it at $40. If the outsourced workflow produced 60 instead of 80, its figure would be $80. Those are arithmetic scenarios, not predicted performance. Costs may also change with volume in a real contract.

Continue tracking the sale and its contribution margin where possible. An appointment that never becomes suitable work is not worth the same as one that does. Use the same period for counting resulting sales and avoid crediting both workflows for the same customer. A small pilot with one extra sale does not establish a permanent conversion advantage.

Before approving a budget, replace every example input with a written quote, measured staff hours or an explicitly labelled assumption. Record low, expected and high outcome cases without presenting the expected case as a guarantee.

Can AI qualify leads as well as a human setter?

Test the specific qualification task rather than making a category-wide claim. Straightforward service-area and scheduling questions differ from advice, negotiation or an ambiguous buying situation. Assess the conversation, the evidence behind the qualification decision and the resulting appointment. A fluent exchange can still lead to an unsuitable booking.

Cost tells you what an appointment took to produce. Next, check whether it was an appointment your team should have accepted.

For an HVAC enquiry, qualification might mean a supported service, an address inside the service area and the right appointment type. A detailed technical question may need a specialist before scheduling.

NIST identifies confident but incorrect AI output as a generative-AI risk and recommends evaluation and monitoring.[1] Apply that to the task: ask about an unapproved price, an unclear service request and a request for a person. This guidance does not establish whether AI or an agency converts better.

Human setters also need accurate information and quality review. Ask either supplier to show why a booking was considered suitable, and agree how disputed appointments are investigated.

When does a combined AI and human workflow make sense?

A combined workflow makes sense when routine enquiries can progress through defined questions but some customers need judgment or a live conversation. Decide the handoff triggers, receiving owner and response commitment in advance. Keep one shared record so the customer does not need to repeat the conversation.

A prospective customer asking for available inspection times may be ready for scheduling. Someone disputing a quote, describing an unusual requirement or asking for a person should enter the appropriate human queue. A helpful handoff includes the original request, answers already given, the unresolved issue and any promise made about timing.

The receiving person might be your employee or a contracted setter. Decide who acknowledges the task, who covers absence and when automation can resume. A notification alone is not a completed handoff. Use the lead follow-up system guide to define ownership and stop conditions across both parts of the workflow.

Proposed handoff design. Verify which steps your chosen providers support and which your team must implement.

What should you ask before signing a software or agency contract?

Before signing, settle the commitments you cannot learn from a demonstration: what is included, what is billable, who handles failures and what happens when you leave. Put those answers in the proposal or contract. The pilot can then test whether the promised service is actually delivered.

For an agency, ask who does the work, which tasks are subcontracted or automated, and what happens when an appointment falls outside your criteria. Clarify credits, replacements, minimum terms and cancellation charges.

For software, confirm setup responsibilities, usage limits, support arrangements and the cost of additional integrations. Ask who updates business information and resolves failures.

For either option, clarify access, retention, deletion, export and incident handling. Agree who owns the phone number and what happens to pending appointments when the service ends. Request a sample report that connects a claimed booking to its enquiry and calendar event.

ACMA states that outsourcing marketing does not transfer away the business’s legal obligations.[2] Include permission records and suppression in the review rather than relying on a general assurance about compliance.

How to run a fair appointment-setting pilot

Use a limited pilot to test delivery and estimate economics. Treat early results as provisional, particularly when the number of attended appointments is small.

  1. Agree scope and success criteria before launch

    Specify lead source, age, channel, offer, service area and qualification criteria. Choose the primary outcome, such as attended qualified appointments per eligible lead. Set acceptable error and handoff standards as well as a budget.

  2. Rehearse the same scenarios with both options

    Test a straightforward booking, an unsuitable lead, an unavailable slot, an unclear answer, a technical question, a human request and an opt-out. Verify the calendar and CRM record, not just the conversation.

  3. Split similar leads between the options

    Where practical, randomly split eligible leads within the same source and time period. Prevent duplicate contact and keep appointment availability comparable. If one option uses calls and the other SMS, report this as a comparison of the complete workflows, not a clean test of AI versus human ability.

  4. Record the work and its actual cost

    Track delivered contact attempts, replies, qualified bookings, attendance, cancellations, opt-outs and complaints. Log setup, review and escalation time. Keep a record of scope or message changes during the pilot.

  5. Wait until appointments have taken place

    Choose a defined group of leads and a reporting cut-off. Show future appointments as pending and report cancellations and reschedules explicitly. Do not compare one provider's completed month with another's first week.

  6. Review quality and decide the next step

    Review a sample from each workflow against the same rubric, including unsuccessful conversations. Pause for failed suppression, false confirmations or unresolved serious complaints. Fix failures, then decide whether to expand, retest or change the division of work.

After the pilot: evidence to use in the decision

Customer fitConversation samples, channel preferences and requests for a person

Decision questionCan customers get the kind of help they need?

Appointment qualityQualification records, calendars and attendee feedback

Decision questionWere the right people booked for the right service?

ReliabilityFailures, duplicate contacts, missed handoffs and recovery records

Decision questionDoes the workflow recover with clear ownership?

EconomicsActual invoices, staff hours and mature appointment outcomes

Decision questionWhat does each useful outcome cost at this volume?

CapacityAvailable slots, response delays and workload at peak demand

Decision questionCan the business fulfil the appointments being created?

ControlSuppression tests, exports, permissions and exit terms

Decision questionCan you inspect, correct and leave the workflow?

DimensionEvidence to inspectDecision question
Customer fitConversation samples, channel preferences and requests for a personCan customers get the kind of help they need?
Appointment qualityQualification records, calendars and attendee feedbackWere the right people booked for the right service?
ReliabilityFailures, duplicate contacts, missed handoffs and recovery recordsDoes the workflow recover with clear ownership?
EconomicsActual invoices, staff hours and mature appointment outcomesWhat does each useful outcome cost at this volume?
CapacityAvailable slots, response delays and workload at peak demandCan the business fulfil the appointments being created?
ControlSuppression tests, exports, permissions and exit termsCan you inspect, correct and leave the workflow?

Agree the importance of each dimension before reviewing results. Failed suppression or fabricated confirmations need resolution even if a provider scores well on price.

What is a sensible decision rule after the pilot?

Choose the option that meets your quality and control requirements at an acceptable total cost and useful volume. Keep the evidence tied to the tested lead source, channel and offer. Expand in stages, because results from a narrow pilot may not transfer to a different audience or more complex conversation.

Use the scorecard to make a specific decision: “This workflow works for these enquiries at this cost.” That is more useful than deciding that one category wins everywhere.

Choose AI where the tested conversations can be completed reliably and your team can own the exceptions. Choose an outsourced service where its people, channels or management provide value you would otherwise have to supply. Keep a combination when the handoff works well for customers and staff.

If neither meets the agreed standard, fix the process or retest before increasing volume. The AI sales-agent buyer’s guide can help if you need to reconsider the type of tool.

Evaluate Vendita against your actual booking workflow

See the SMS qualification and calendar-booking process, then compare the scope, operating requirements and cost with your other options.

Explore Vendita appointment booking