"Is it working" is the wrong question to ask about an AI receptionist, because the honest answer is always some version of "sort of, probably." The right question is a number: what does it cost you, per appointment actually booked. Most businesses either don't calculate this or calculate it in a way that doesn't hold up. Here's how to do it properly.
The formula
It's simpler than people expect:
Cost per booked appointment = total AIVA spend in a period ÷ number of appointments actually created through AIVA in that period.
It's a close cousin of what marketers call customer acquisition cost — total spend divided by outcomes — just scoped to a single channel and a single kind of outcome instead of a whole business's marketing budget.
The two words that matter are "actually created." Not conversations. Not calls answered. Not even calls that sounded like they were going well. Appointments that landed in your calendar or booking system. Everything else is a leading indicator, not the number.
Why a flat monthly fee makes this number meaningless
If you were paying a flat monthly fee for a receptionist tool, this formula still technically works, but it tells you almost nothing useful. In a slow month with few bookings, you divide a fixed number by a small denominator and the cost per booking spikes — not because anything got worse, just because volume dropped. In a busy month, the same fixed numerator divided by a bigger denominator makes the cost per booking look great — again, not because anything improved, just because volume happened to be high. The metric swings with volume instead of tracking performance, which makes it useless for the one thing you actually want to use it for: deciding whether this is worth what you're paying.
Why usage-based pricing fixes this
On usage-based pricing, the numerator moves with the denominator. If bookings double, spend roughly doubles too — you're paying for twice the conversations, not a flat fee stretched thinner. That means cost per booking stays a genuinely stable number across slow and busy months, and a real change in it, up or down, is telling you something true about how well AIVA is converting conversations into bookings, not just reflecting how busy the month was.
A worked example
Say your dashboard shows, for the month: ₹6,000 spent on voice, ₹1,400 on web chat, ₹300 on SMS. Total spend: ₹7,700. Your booking system shows 165 new appointments that came in through an AIVA conversation that month.
₹7,700 ÷ 165 = ₹46.67 per booked appointment.
Once you have a real cost-per-booking number, the next question answers itself: what's a booked appointment actually worth to you? If the answer is anywhere near ₹500, you're not looking at a cost. You're looking at a margin.
A second example, at a smaller scale
The math holds the same way at a very different size. Say a solo consultant running a much lighter setup spends ₹1,200 in a month — mostly web chat and a handful of voice minutes — and gets 18 bookings out of it. ₹1,200 ÷ 18 = ₹66.67 per booking, noticeably higher than the ₹46.67 example above. That's not automatically a worse result. A lower-volume account often has a higher share of exploratory, non-booking conversations relative to its total spend — someone asking a pricing question who isn't ready to book yet still costs something to answer, even though they don't show up in the denominator. The number only means something next to what a booking is actually worth to that specific business, which is a separate figure you have to bring to it, not something the formula produces on its own.
A third example: a seasonal business
The math holds up under a swing in volume too, which is worth checking directly since that's exactly the situation where a flat monthly fee would have distorted the number badly. Say a wedding catering business spends ₹2,200 in a slow month — October, say — and gets 12 bookings out of it: ₹183 per booking. In peak wedding season a few months later, spend jumps to ₹9,000 across a much busier month, and bookings jump to 58: ₹155 per booking.
Notice what didn't happen: the number didn't swing wildly between the quiet month and the busy one, even though spend nearly quadrupled. That's usage-based pricing doing its job — a big month costs more in absolute terms, but the cost per outcome stays in a broadly comparable range, which is exactly what you want from a metric you're using to judge performance rather than just measure volume. A business with a genuinely seasonal calendar — caterers, decorators, anything tied to festival or wedding timing — should expect the raw spend number to move a lot month to month. The cost-per-booking number moving only a little in comparison is the reassuring sign, not the concerning one, and it's worth checking a full season's worth of months before drawing any conclusion from a single one.
The zero-booking edge case
It's worth naming a genuine edge case directly: what happens if a period has real AIVA spend but zero completed bookings — a brand-new account still in its first days, or a business whose booking system integration isn't fully wired up yet. The formula technically breaks (you can't divide by zero), but the more useful response isn't to force a number out of it. It's to treat a zero-booking period as its own signal — either the volume sample is too small to mean anything yet, or something upstream, like a calendar connection, needs checking before the cost-per-booking metric is worth trusting at all. Don't average a zero-booking week into a monthly figure and call the result meaningful; widen the window until you have enough real bookings for the denominator to be doing real work.
Do cancellations and no-shows count?
A booking that gets cancelled later, or turns into a no-show, still counts as a booking in the period it was created — the formula measures whether AIVA successfully turned a conversation into an appointment, not whether that appointment was ultimately kept. That said, cancellation and no-show rates are worth tracking as their own separate numbers alongside cost per booking, not folded into it. If your cancellation rate is unusually high specifically on AIVA-booked appointments compared to appointments booked another way, that's a distinct problem worth investigating on its own — possibly in your confirmation and reminder flow — rather than something that should quietly deflate your booking count after the fact.
Splitting it by channel
Because voice, chat, and SMS are priced and tracked separately, you can run this same formula per channel and see which one is actually converting efficiently for your business. Some businesses find web chat turns browsers into bookings cheaper than voice does; others find the opposite, because their customers trust a phone call more than a chat window. Neither is right in general. It's specific to what your customers do. If you're chasing this number down to decide where to invest more setup effort — better FAQs, a more detailed booking flow — the channel split is usually the more actionable version of the metric, since it points at a specific conversation type instead of an average across all of them.
What a booking is actually worth, beyond the first visit
The comparison in the pull-quote above uses a single visit's value, which is the simplest honest starting point. But for a lot of businesses — clinics with repeat visits, salons with a regular client base — the real comparison should use something closer to lifetime value than a single ticket, since a booking that becomes a returning customer is worth considerably more than its first appointment alone. We've walked through the fuller version of this math for two specific business types in what a clinic can save in a year and what a salon can save in a year, both of which build outward from the same per-booking logic covered here.
Benchmarking against what you used to pay
Cost per booking is also worth comparing against the thing it's most often replacing: a person answering the phone. Take a rough, honest version of that math — a receptionist earning ₹18,000 a month, handling, say, 220 bookings across that month alongside everything else already on their plate. That's roughly ₹82 per booking in salary cost alone, before adding the phone line, the desk, or the hours they're unavailable during lunch, after closing, or on the days they're out sick and nobody's covering the phone at all.
That's not an argument against hiring a receptionist — plenty of businesses need one for reasons that have nothing to do with phone volume. But it's a useful sanity check on the AIVA number specifically: if your cost per booking lands meaningfully below what a person handling the same volume would cost per booking, that's the comparison actually worth making, rather than treating any nonzero cost per booking as money that wouldn't otherwise have been spent somewhere. The honest version of "is this worth it" is never AIVA against free — it's AIVA against whatever was answering the phone before, including all the hours nobody was answering it at all, which is where a lot of the real savings actually live.
A common mistake: comparing your number to someone else's
It's tempting to ask around and benchmark your cost per booking against another business's, and it's usually the wrong comparison to make. A ₹47 cost per booking at a clinic booking ₹1,500 consultations and a ₹47 cost per booking at a boutique booking ₹8,000 custom orders are not telling you the same thing, even though the number matches exactly — the second business is getting vastly more value per rupee spent, because what's being booked is worth more to begin with.
The comparison that actually means something is your own number against your own average booking value, and your own number this month against your own number last month or last quarter. A rising cost per booking at your business, measured against your business's own recent history, is a real signal worth investigating. The same number sitting next to a different business's number, in a different category with a different average ticket, mostly isn't — it just happens to look alike on a spreadsheet.
Using it to test changes, not just to monitor
Cost per booking is also useful as a before-and-after measure, not just an ongoing dashboard number. If you update AIVA's FAQ answers, adjust your booking flow, or add a new escalation rule, compare cost per booking for a comparable window before and after the change. A real improvement in how well AIVA converts conversations into bookings should show up here specifically, separate from whatever's happening with overall call volume that month — which is exactly the distinction a flat monthly fee could never let you make cleanly.
What to watch, not just calculate
Your analytics dashboard tracks resolution rate alongside your spend, which is worth watching next to cost per booking, not instead of it. A rising cost per booking with a falling resolution rate tells you something is going wrong in the conversation itself, worth investigating before you touch pricing at all. A rising cost per booking with a steady resolution rate is more often just seasonal demand shifting the mix of callers — worth knowing about, but not a sign anything's broken, especially if your business runs a genuinely seasonal call pattern to begin with.
Calculate it monthly, keep the channel split, and you'll have an actual answer to "is this working": a number, not a feeling. The pricing page has the underlying per-unit rates if you want to sanity-check your own spend total against your logged call and message volume directly.