Dental Clinic AI Receptionist India: ₹2/min

Dvaarik is self-serve SaaS today: there is no setup fee and no minimum commitment. You log in, add your services, staff, hours and knowledge, connect channels, and pay only for what your agents actually do.

Credits are prepaid and never expire; at zero balance agents pause gracefully, so there are no surprise bills. Agencies and developers can sell the same platform under their own brand through the white-label partner programme, and set their own client pricing.

We will only quote prices we read on the vendor's own page during this task. Everything else stays out, including several figures floating around in search results that we could not confirm.

Note on like-for-like: both Aixclerate plans include 1 business number in the plan price. Ours is charged separately at ₹1,500 per 30 days, which is why the Dvaarik row shows the blended figures rather than only the ₹2.00 talk-time rate. If you point an existing clinic number at us, that charge disappears and the rate is ₹2.00 flat.

Aixclerate plan contents and add-on prices verified on aixclerate.com/pricing, checked 20 Jul 2026: Starter includes 2 AI agents, 2 concurrent calls and 1 business number; Business includes 4 agents, 4 concurrent calls and 1 number; extra business number ₹699 (withdrawn 29 Jul 2026), call recording ₹0.20/min, additional concurrency ₹499 (withdrawn 29 Jul 2026); annual billing advertises 2 months free. Their page states no setup fee and no GST treatment. The ₹20.00 and ₹12.50 figures are our arithmetic — plan price divided by included minutes — not their claims.

Caller Digital's absence of an entry rate verified on caller.digital/voice-ai-pricing-india, checked 20 Jul 2026; the same page states its outcome-based tier is only available for deployments above 100,000 minutes per month.

The category range, and the gap between headline and reality

Caller Digital describes India voice AI headline rates in 2026 as "₹2–₹12 headline" with "₹3–₹6 is the most common mid-market range" (verified on caller.digital/voice-ai-pricing-india, checked 20 Jul 2026). The same page argues that effective cost runs "2–4× headline once you add platform/SaaS fees, telephony markup over TRAI rates, and connect-rate losses. A ₹3/min quote typically costs ₹6–₹9/min in production" (same source, same date, exact quote).

We think that critique is largely fair, and it is worth applying to us too. So here is the multiplier check on Dvaarik, honestly:

• Platform/SaaS fee: none. There is no monthly plan to add on top.
• Telephony markup: the ₹1,500 per 30 days number charge is the telephony line, stated separately rather than buried in a per-minute rate. At 480 billed minutes a month that ₹2,000 is effectively ₹4.17/min on top, taking your true blended rate to about ₹6.17/min. At 3,000 minutes it is ₹0.67/min, blended ₹2.67. Our model gets cheaper per minute the more you use it, and is genuinely expensive at very low volume. If you already have a working number you want us to answer on, that line disappears entirely.
• Voice tier: ₹2/min is Bharat Standard, which covers the 22 scheduled Indian languages and English. Studio HD is ₹3/min and Global voices for non-Indian languages are ₹4–5/min. If your call mix is entirely Indian-language, ₹2/min is what you pay.
• Connect-rate losses: real for outbound recall campaigns. Unanswered ringing is not billed as conversation, but every connected call that ends in "call me later" is still billed talk time. Budget for it.
• Whole-minute rounding: if your calls average 45 seconds — plausible for a simple "are you open on Sunday" enquiry — you pay for 60. That is a real ~33% markup on short calls versus per-second billing, and we do not pretend otherwise. Longer booking conversations round much less painfully.

Caller Digital also publishes per-call pricing at "₹4–₹15 per connected dispositioned call" and outcome-based pricing at "₹8–₹25 per successful outcome" as market structures (same source, checked 20 Jul 2026). If a vendor offers you per-outcome pricing for a dental clinic, ask precisely what counts as an outcome and who adjudicates a disputed one. See also AI calling agent price in India 2026.

Numbers we refused to publish

While researching this page we found widely circulated claims about dental no-show reduction percentages, monthly revenue recovered by Indian clinics, and the share of calls dental practices miss. Every one of them traced back to a search snippet whose source page we could not open or which cited nothing. They are not in this article. If a vendor quotes you a no-show reduction percentage, ask which clinics, over what period, measured how.

The subject described in this section: This is the comparison most owners will want to run, so let us do it with real published salary data rather than a convenient guess.

Treat both cautiously. Ten reported salaries is a very small sample and self-reported salary data skews toward metro and organised employers. Your actual cost in Hyderabad differs from Paloncha. We publish the sample size because a figure without one is not evidence.

The comparison, framed honestly

The honest conclusion is not "fire your receptionist." It is that these two things are bad substitutes and good complements. Your front-desk person is most valuable with the patient standing in front of them. The AI is most valuable at 9:40pm, during the lunch rush when both lines ring, and on the recall list nobody will ever get around to working.

The pressure to answer every call is structural in Indian dentistry. India has 376,721 registered dentists — roughly one dentist per 3,846 people, against a WHO recommendation of 1:7,500 for developing countries — with 329 dental colleges and 27,926 BDS seats a year (Observer Research Foundation, Damini Bisen, "Dentistry in India: Oversupply and Gender Bias Shaping Oral Health Workforce", 21 Nov 2025, attributing the figures to dciindia.gov.in; checked 20 Jul 2026). In a market with that density, a caller who hits voicemail has another clinic a short scroll away. We are not going to attach a percentage to that, because we do not have one.

Described mechanically, with no performance claims attached. All examples below use invented clinic and patient names.

1. Answering and booking

The agent answers every call, including outside clinic hours, identifies the practice by name, and works through a booking script you approve. It offers real open slots, holds the slot while it confirms details, then writes the appointment and sends a WhatsApp confirmation.

Illustrative exchange (invented clinic, "Sunrise Dental, Kukatpally"):

> Caller: I want to come for cleaning, do you have Saturday. > Agent: This is Sunrise Dental. I can offer Saturday 11:30am or 4:15pm for a scaling appointment with Dr Menon. Which suits you better. > Caller: 4:15. > Agent: Holding 4:15pm Saturday. May I have your name and the number you would like the confirmation sent to.

Slot-locking matters more in dentistry than in most verticals, because chair time is the constraint. Details on booking.

2. Reschedule and cancel — with the slot released

A cancellation that nobody records is a chair sitting empty. The agent handles reschedules and cancellations on the same call and releases the slot back into availability immediately, which means it is offerable to the next caller rather than discovered as a gap on the morning list.

3. Six-month recall calls

The subject described in this section: This is a workflow that takes hands most front desks do not have spare. You supply the recall list from your records. The agent calls each patient during the calling hours you set, identifies the practice, states why it is calling, and either books a slot or logs the outcome:

• Booked — appointment written and confirmation sent
• Call back later — rescheduled attempt per your rules
• Not interested — flagged and not called again
• No answer — retried per your retry policy

See outbound calls for how campaigns are structured.

4. Treatment-plan follow-up

A common pattern in dentistry, as we understand it: a patient accepts a root canal, crown or aligner plan at the chair, then leaves and never books the next stage. You give the agent that list; it calls, references the accepted plan in general terms only, and offers slots. It does not discuss the clinical merits of the treatment, restate the diagnosis, or answer "do I really need it" — that goes back to the dentist.

5. Deposits and payment links

For high-value treatment where a no-show costs an hour of chair time, the agent can send a payment link mid-conversation and confirm the booking on receipt. No commission is taken on what you collect. See payments.

6. Reminders that actually get read

Reminders go out on WhatsApp, on your own number, with a voice reminder call as an option for patients who do not use WhatsApp. WhatsApp message fees are billed by Meta to your own Meta account at zero markup from us — see WhatsApp Business API costs in India.

We are not going to tell you what this does to your no-show rate. We do not have that number for dental clinics in India and we are not going to borrow someone else's unsourced one. Measure it in your own clinic over a quarter; that is the only figure worth having.

The subject described in this section: It is the obvious second question after price, and it is the one we get asked to answer in writing. Here is our position.

India's Telemedicine Practice Guidelines — framed for registered medical practitioners under the Indian Medical Council Act, 1956 — prohibit AI and machine-learning platforms from counselling or prescribing. The guidelines were issued on 25 Mar 2020 by the Ministry of Health and Family Welfare and appended to the Indian Medical Council (Professional Conduct, Etiquette and Ethics) Regulations, 2002. The relevant wording: *"artificial intelligence and machine learning based platforms cannot counsel or prescribe any medicine to any patient"*, and such technology *"can only be used to aid and support the RMP, but the final advice and prescription must be made by the RMP"* (Intelehealth, author Nikhil Issar of PSA, "Medicolegal implications of India's new telemedicine guidelines", posted 24 Apr 2020 and updated 10 Mar 2022; checked 20 Jul 2026).

The same restriction is confirmed in the peer-reviewed literature: *"The use of artificial intelligence or machine learning applications to counsel or to prescribe medicines is strictly prohibited"* (Damodharan D, Manjunatha N, Kumar CN, Math SB, "Telemedicine practice guidelines of India, 2020: Implications and challenges", Indian Journal of Psychiatry, 2021; checked 20 Jul 2026).

Whether those guidelines bind dental practice specifically is not something we can state. Dentists in India are registered under a separate statutory regime, and neither source we could read addresses dental practice. We apply the same boundary regardless, by choice. A note on sourcing: the MoHFW's own PDF was not retrievable while writing this, so both citations above are secondary sources. Confirm the current text with your own legal adviser before relying on it — regulations change and we are not lawyers.

What that means for the agent, concretely

The agent applies a routing rule, not a clinical assessment. It listens for urgency language — severe pain, swelling, bleeding that will not stop, a knocked-out tooth, injury after an accident — and on hearing it, stops trying to book and escalates.

Illustrative handling (invented patient scenario):

> Caller: My tooth broke and there is a lot of bleeding, what should I do. > Agent: I am not able to give any medical advice. I am connecting you to Dr Menon's emergency line right now. If the call does not connect, I have your number and the doctor will call you back.

The boundary is a configuration you sign off on, not a setting we choose for you. You decide the escalation number, the after-hours rule, and the exact wording of the refusal. What is not configurable is whether the agent can give clinical advice. It cannot, for any clinic, under any configuration.

Dental patients in India do not speak in one language per call. They open in English, switch to Telugu to explain where it hurts, and say "appointment" and "cleaning" in English regardless.

The agent handles all 22 scheduled Indian languages plus code-mixed Hinglish, Tenglish and Tanglish, and most world languages, with mid-sentence switching rather than a menu choice at the start of the call.

The practical contrast is with a keypad IVR, where "press 1 for Hindi" is where most patients over 60 hang up. More in multilingual AI receptionist for Hindi and Telugu, multilingual and AI voice agent vs IVR.

Will elderly patients talk to it. Some will, some will ask for a person immediately. That is why the transfer-to-human path is configured on day one and why the agent never argues with a caller who asks for a human. We do not have data on Indian dental patients by age group and will not invent any.

Dental records are sensitive. Here is our position, limited strictly to what we can actually commit to.

Outbound conduct rules, non-negotiable

• We call only your own patients and enquiries. No purchased lists, no scraped numbers, no cold databases. Ever.
• The agent identifies your practice by name at the start of every outbound call.
• Calling hours are respected and set by you. No recall calls at 8am on a Sunday unless you configure it.
• A patient who says stop is marked and not called again by that campaign.

Data handling

• The list is yours — you load it, we do not sell, share or reuse it.
• Call recording is optional and priced at +₹0.10/min when enabled. If you do not enable it, calls are not recorded.
• Transcripts and call outcomes are available to you.

What we will not claim

Several competitor pages carry a compliance section citing the DPDP Act 2023 with section numbers. We tried to read the Act's official text while writing this and the government source was not retrievable, so we are not going to quote sections we did not read. What we will say is the operational position: consent for calling and messaging sits with the practice as the party holding the patient relationship; take your own legal advice on this, and on retention, before running a recall campaign at scale. Any vendor confidently reciting the Act at you should be asked whether their platform architecture actually implements what they are quoting.

We also make no specific data-residency, retention-window or certification claim on this page. If that is a procurement requirement for you, ask us directly at contact and we will answer honestly, including where the answer is no.

We build the agent for you. This is a studio model, not self-serve software you configure yourself — there is no dashboard to learn, no plan to pick and no trial to start.

What you provide

1. Your services and prices — consultation, scaling, RCT, crown, aligners, whatever you offer, with what you are willing to quote on a call and what must be quoted in person.
2. Chair availability — working hours, dentist-wise schedule, slot lengths per procedure type, holidays.
3. The escalation path — which number an emergency routes to, in and out of hours, and the fallback if nobody answers.
4. Your booking destination — Google Calendar or a shared calendar we can write into.
5. Recall and follow-up lists, if you want outbound running from week one.
6. The 20 questions your desk answers daily — parking, insurance, do you do painless extraction, is Dr X in on Thursday.
7. Sign-off on the emergency refusal wording.

What we do

Live in days, not months, for a single-location clinic. Every agent is built personally by me — Rohith Sriramula, a laid-off engineer working from Paloncha in Bhadradri Kothagudem, Telangana. There is no team, no office and no support desk, and the next section explains when that should send you elsewhere. See how it works and about.

Before committing, hear one. The live demo is the only proof we have and we would rather you judge it than take our word: try it from the pilot page.

Dvaarik is self-serve SaaS today: there is no setup fee and no minimum commitment. You log in, add your services, staff, hours and knowledge, connect channels, and pay only for what your agents actually do.

Credits are prepaid and never expire; at zero balance agents pause gracefully, so there are no surprise bills. Agencies and developers can sell the same platform under their own brand through the white-label partner programme, and set their own client pricing.

Four situations where a competitor genuinely serves a dental clinic better than we do. This section is here because the rest of the page is only credible if this one is honest.

1. You run dental practice-management software

The subject described in this section: This is our biggest real gap. Dedicated dental products integrate natively with dental PMS and write appointments straight into the chair-side schedule alongside treatment history and recall flags. We have no dental-PMS integration. Our booking is calendar and WhatsApp based with slot-locking, which fits the very large number of Indian dental clinics running on a paper diary, a WhatsApp group or Google Calendar — and fits poorly if you already have a system your staff live inside all day. If a nurse would have to copy appointments from our calendar into your PMS by hand, we are creating work, not removing it. Choose a PMS-integrated vendor.

2. You are a multi-branch group with heavy simultaneous call volume

Our default is one number and one concurrent line, with extra lines at ₹1,500 per 30 days each. That scales to a few lines, not to a call centre. Aixclerate sells 4 concurrent calls in a published plan; Caller Digital's page describes serving deployments above 100,000 minutes a month (both checked 20 Jul 2026). If you need real concurrency at scale, an SLA and a support rota, buy from a company that has those. We are one engineer.

3. You need to see proof before you buy

Every competitor page shows testimonials, case studies and outcome percentages. Today Dvaarik is self-serve with no setup fee and no required plan — usage is prepaid credits, voice from ₹2/min with whole-minute billing. If your decision requires evidence that it worked at a comparable clinic, we cannot give you that today. The only honest counter we have is that you can call the live demo yourself and judge the thing directly rather than judge a screenshot of a testimonial. That is a real disadvantage, and we would rather you hear it from us.

4. You want one round monthly number and no arithmetic

Our model wins on total cost at most realistic clinic volumes — the worked scenarios above show the crossover — but it loses badly on cognitive simplicity. If your practice manager wants a fixed line item and no variability, a flat plan is a legitimate choice even at a higher effective rate. Predictability has value.

If none of those four apply, the case for us is straightforward: a published rate, no voice plan, and a boundary on medical advice we will not cross for anyone.

Take this list to us and to everyone else. The answers are more revealing than the demos.

1. What is your per-minute rate, in rupees, published on a page I can read. If the answer is "book a demo", note that they wrote a pricing page without a price on it.
2. Is billing per second or per whole minute. Whole-minute rounding costs you materially if your calls are short. Ours is whole-minute, and we say so.
3. What do I pay in month one before a single call is answered. Setup, plan fee, number, minimum load, onboarding.
4. Am I billed for ringing, unanswered calls and voicemail, or only connected conversation.
5. What exactly happens when a patient describes severe pain at 2am. Ask for the literal script; our agent gives no clinical advice of any kind, and you should be able to read any vendor's refusal wording before you buy.
6. Does the agent write into my dental practice-management system, or into a calendar. Ask for the specific system name.
7. How many calls can it answer at once, and what does each additional line cost.
8. Where does the WhatsApp message cost land — on your bill with a markup, or on my own Meta account.
9. Is there a lock-in, a notice period, or do credits expire.
10. Every statistic on your website — which clinic, what period, measured how. If it cannot be sourced, discount the page.
11. Who builds it and who fixes it at 9pm when it mishandles a call. A named person is a better answer than a support tier.
12. Will you call numbers I did not give you. The only acceptable answer is no.

Related reading: best AI voice agent platforms in India 2026, cloud telephony vs AI voice agent, and the FAQ.