Is an AI Receptionist Worth It? The ₹2,600 Test

Yes, if you are losing enquiry calls at a rate a person cannot fix. Nights, Sundays, lunch hour, and the second caller arriving while the first is still on the line. Those calls are not a staffing failure. One person holds one conversation at a time, and a shift is eight hours out of twenty-four.

No, if your call volume is very low. The ₹2,000 phone number recharge every 30 days is a fixed cost that does not shrink when your usage does. At 100 answered minutes a month it works out to about ₹67 per answered call. At 300 minutes it is about ₹26. At 1,000 minutes it is about ₹12. Below roughly 200 answered minutes a month you are mostly paying for a phone number, and a person returning calls at 6pm is the cheaper answer. If you still want something automated at that volume, a vendor that bills purely per minute with no monthly floor is a better-shaped product than ours — ask what their phone number costs before you believe the rate.

Maybe, and it is decidable in twenty minutes. Count seven days of missed calls in your phone log, multiply by four, apply the conversion rate you already know, and compare against the tables below. If the arithmetic is close, it is a no — a marginal case is not worth the disruption of changing how your phone works.

It means recovered profit exceeds monthly cost. Cost we publish, so that side is settled. The recovered side has three inputs, and all three are yours.

1. Missed enquiry calls per month. Not total missed calls — enquiry calls. Your log will also hold suppliers, delivery riders, wrong numbers, spam, and existing customers asking something that could wait until tomorrow. None of those are worth ₹26 to answer at 2am.

2. The share of those callers who would never have come back. Plenty of people ring again in the afternoon, or send a WhatsApp, or walk in anyway. A missed call is only a lost customer if that particular caller went elsewhere and did not return. Any break-even sum that skips this step has quietly set it to 100%, and it is never 100%. If you have no idea what the share is for your business, assume half and see whether the arithmetic still works — if it only works at 100%, it does not work.

3. What you keep on one converted customer. Profit, not invoice value. A ₹40,000 dental implant and a ₹400 haircut sit at opposite ends of this page, and they get opposite answers at the same call volume.

Multiply those three together and you have monthly recovered profit. Compare it to the monthly cost in the next table. That comparison is the entire decision, and no vendor — including us — can do it for you, because we do not have your call log.

Dvaarik's own published rates, at volumes a small business would recognise. Voice is ₹2 a minute from the first minute with simple per-minute billing. The ₹2,000 phone number recharge every 30 days comes out of the same prepaid credit balance. The call count assumes a three-minute average call, which is an assumption about your business, not a Dvaarik figure — a clinic booking runs shorter, a property enquiry longer.

Read the last column first, because it is the one that decides whether this is worth it. The per-minute rate is identical in every row. The cost per call falls by more than seven times across the table, entirely because a fixed ₹2,000 is being spread over more calls. Anybody selling you on a per-minute rate at 100 minutes a month is selling you the wrong number.

Three things that table leaves out, stated rather than buried. GST: all our prices exclude GST and 18% is added when you load credits, so ₹2,600 of usage is billed ₹3,068 — whether that GST is recoverable depends on your own registration, which is a question for your accountant. Months one to three for founding clients: the Care Plan at the founding rate of ₹5,000 a month is required for the first three months and is debited from the same credit balance, so add ₹5,000 to every row for those three months — 300 minutes becomes ₹7,600 ex-GST, or about ₹76 a call. A second concurrent line, so two callers can be answered at once, is ₹1,500 a month and is not in the table.

Take the ₹2,600 row — 300 answered minutes a month, month four onward, before GST — and divide it by what one converted customer leaves you. Months one to three are shown alongside at ₹7,600, because for founding clients the Care Plan is required for those three months and comes out of the same balance, so the break-even is genuinely higher while it does.

The ₹25,000 row is why a dental clinic, a property desk and a banquet hall reach a different verdict from a tiffin service at the same call volume. One recovered implant case, one recovered site visit, one recovered wedding booking pays for most of a year. At the other end, a ₹500-margin business needs six genuinely recovered customers every month, every month, and should be sceptical that it will get them.

And the word doing the work is recovered. Six recovered customers is not six answered calls. If a third of the callers you currently miss would have rung back anyway, and one in three enquiries converts, six recovered customers means roughly twenty-seven missed enquiry calls a month — which is about one a day. That is a real business with a real problem, not a hypothetical one. But it is also a number you should confirm in your own call log rather than take from a vendor's page.

Six situations. We would rather tell you now than three months in.

Your call volume is very low. Under roughly 200 answered minutes a month, the fixed ₹2,000 recharge dominates and you are paying about ₹36 or more per answered call. A shop taking five calls a day and answering four of them is not losing enough to fund this. Return the missed one at 6pm.

You already answer nearly everything. Check before you assume. If your missed-call rate is a few percent and those callers ring back, there is almost nothing to recover, and a page like this one should not talk you into a purchase.

The missed calls are not buyers. If the log is mostly suppliers, couriers and spam, answering them faster earns nothing. Sort the misses by who was calling before you sort them by count.

The job is the counter, not the phone. Walk-ins, documents, cash, keys, showing someone where to sit — no software on this page does any of that, ours included. If the person you are picturing is standing at a desk in front of customers, hire the person.

Your bottleneck is downstream. If your calendar is already full, or nobody follows up on the enquiries you do capture, more answered calls just produces more people you disappoint. Fix the follow-up first; it is free.

You want it to sell to strangers. Our agents call your own enquiries and your own customers — people who contacted you. We do not call bought lists or cold databases, so if the plan was outbound to a purchased database, we are the wrong supplier and you should say so early.

Usually not, and the reason is the same fixed-cost effect. At small-business volume, what mostly decides your monthly bill is what you have to pay before you place a single call — and for most vendors below that is a monthly plan rather than a rate. Every figure in the table was read on the vendor's own page on 21 Jul 2026.

That fixed-floor column is also why the table does not settle the very-low-volume case in our favour. A vendor billing purely per minute, with no monthly plan underneath, gets cheaper than us as volume falls — Botsense, for instance, publishes ₹9/₹7/₹5 a minute and no monthly floor on that page, though it publishes no phone number cost either, so what you would actually pay them per month is not established.

Now the honest part. On published rates alone, 300 minutes is ₹2,700 of talk-time at Botsense's ₹9/min Starter rate against ₹600 at ours — but their page publishes no phone number cost, so ₹2,700 is not a complete monthly bill and neither, in isolation, is our ₹600. That arithmetic is ours, derived from their published rate, not a figure Botsense states. It is also the clearest illustration on this page of why shopping the per-minute rate is a mistake: at 300 minutes a month, our all-in ₹2,600 and their ₹2,700 of talk-time are close enough that the rate is not what should decide it.

We are not claiming to be the cheapest, and you should distrust anyone who does. The lowest headline number here, ₹0.40/min, is one Scalify's own page labels an estimate for which you should always get a formal quote; on their service page voice minutes are billed separately on top of a plan starting ₹15,000 a month; and their stated 10,000-minute monthly minimum puts it out of reach of a business doing 300 minutes anyway. What we will claim is the terms: ₹2 a minute from the very first minute, no subscription, no minimum volume, no lock-in. For a business that is genuinely unsure whether this is worth it, terms matter more than rates, because terms are what let you stop.

Four things, and the first one is the big one.

Not every missed call was a lost customer. People ring back. They WhatsApp. They walk in on Saturday anyway. Any calculation that treats every missed call as a lost sale overstates the return by however large that share is — including the quick version in our own calculator, which never asks you the question. Halve your estimate and see whether it still clears.

Answering does not create demand. This recovers calls you were already getting. If enquiries are down because the ads stopped or the season turned, the receptionist is not the lever.

Some callers want a person, and should get one. A well-built agent hands over rather than improvises, which is the correct behaviour and also means some of the calls it answers still end up needing your team. We wrote separately about how Indian callers actually react to this, including where they do not like it, in do customers accept AI answering calls.

It is only as good as what you told it. An agent that does not know your Sunday timings, your price list or which doctor is on leave will answer confidently and wrongly. That is what the build week and the Care Plan tuning are for, and it is why we build each agent for the specific business rather than handing over a dashboard.

Do not decide this from a page written by a vendor, including this one. Do this instead.

1. Open your call log and count seven days. Every inbound call, whether it was answered, and roughly what the caller wanted. Ten minutes on a phone.
2. Keep only the enquiry calls that went unanswered. Drop suppliers, couriers, spam and existing customers with a question that could wait.
3. Multiply by four for a monthly figure.
4. Apply the share you believe would never have come back — if you do not know, use half.
5. Multiply by your conversion rate, the one you already know from experience.
6. Multiply by the profit you keep on one customer.

That is your monthly recovered profit. Compare it to the row in the cost table that matches your volume. If it clears the cost with room to spare, it is worth it. If it is close, it is a no — a marginal purchase is not worth changing how your phone works.

Our missed-call calculator will do a rougher version of this from your own numbers in a few seconds, free, with no email and no form. Three things it does differently, so you can discount its answer accordingly: it works from your average ticket value rather than the profit you keep, it does not ask whether those callers would have rung back, and the cost it shows is talk-time only, without the ₹2,000 recharge. It lands on the optimistic side. The six steps above are the honest version, and it is the only tool on our site we would be happy for you to use and then not contact us.

You tell us what you are losing, we tell you what we would build, and we build it — days rather than months, with the founder building it personally.

How it sits in front of your existing number depends on your own operator, so we check that before promising anything. Where conditional call forwarding is available on your line, calls ring your phone first exactly as they do now and only the unanswered ones reach the agent. Where your operator does not support it, the agent answers on the number that comes with your recharge and you point your ads and listings at that. Either way, what goes unanswered — after closing, on a Sunday, while your one line is busy — gets answered, in the caller's language across 22+ Indian languages including code-mixed speech, and a summary drops into a sheet or on WhatsApp so your team sees it in the morning. Anything sensitive gets handed to a person.

The money, in full. Voice is ₹2 a minute, whole-minute billing, permanent — not a launch discount. Chat is ₹2 per conversation across a 24-hour window. The minimum first credit load is ₹10,000, billed ₹11,800 because 18% GST is added when you load credits, and month one debits ₹2,000 for the number and ₹5,000 for the Care Plan at the founding rate, leaving roughly ₹3,000 of usable credit. Top-ups later start at ₹500, credits never expire, and at zero balance everything pauses gracefully rather than running up a bill. A custom build is ₹40,000 and is free for our 10 founding clients — a trade stated plainly rather than in fine print: the free build is in return for taking the Care Plan for the first three months at the founding rate of ₹5,000 a month, and giving us an honest testimonial and a short case study after go-live. From month four the Care Plan is optional.

Before any of that, hear it. The live demo agent is the same technology we would build on, and it costs you nothing to argue with it for five minutes. Then message us on WhatsApp at +91 93923 98750 with your call volume, and we will work out your real monthly number — including telling you not to buy, if your volume does not support it.