Overview
The group runs 12 dental clinics across Chennai, covering general dentistry, orthodontics, implants and cosmetic work, with around 95 staff and 40 chairs.
The economics of dentistry hinge on treatment-plan conversion. A patient consults, accepts a multi-visit plan, perhaps a root canal followed by a crown, but then delays scheduling the next visit. Each lapsed plan is recurring revenue and a clinical outcome left incomplete.
In early 2026, the clinical director found that nearly 45% of accepted treatment plans went unscheduled beyond 30 days, and that front-desk staff had no consistent way to chase them. They wanted an agent that could follow up on every accepted-but-unscheduled plan in the patient's language, book the next visit on the right dentist's chair, and log the outcome in the practice-management system.
The challenge
The follow-up gap had four compounding failure modes: no systematic chase, language friction, no-show leakage, and no clean conversion data per clinic.
- 45% of accepted treatment plans went unscheduled. Patients accepted multi-visit plans at consultation but never booked the next step, and the front desk had no reliable process to follow up across 12 clinics.
- Tamil-first patients disengaged from English calls. Around 60% of patients preferred Tamil, and English-default scripts saw far higher early hang-ups, especially among older patients needing implants and dentures.
- 33% no-show rate on booked procedures. Without structured reminders, a third of booked procedures were missed, leaving high-value chair time idle and pushing back other patients.
- No per-clinic treatment-conversion visibility. Conversion from accepted plan to completed procedure was tracked inconsistently, so the group could not see which clinics or dentists had the biggest follow-up gaps.
- Front desk could not chase and serve walk-ins at once. Reception staff were stretched between in-clinic patients and follow-up calls, so the follow-up calls almost always lost.
The AI-powered solution
Kallix deployed an AI voice agent named Priya, a warm Tamil-English voice, that reads accepted-but-unscheduled treatment plans from the practice-management system, follows up in the patient's language, books the next visit, and writes the outcome back. The full build took 14 working days.
Daily lapsed-plan queue from the PMS
Each morning the agent pulls accepted plans with no scheduled next visit beyond a set threshold, prioritising higher-value and clinically time-sensitive procedures.
Tamil / English detection and switching
The agent meets each patient in their preferred language, switching mid-call when patients code-switch, which kept older implant and denture patients engaged.
Procedure-aware booking
The agent books the right chair and dentist for the specific procedure, respecting slot lengths for cleanings, root canals, crowns and implant stages.
Cost and prep clarity on the call
The agent confirms the next-step cost, any pre-procedure prep, and answers common questions, removing the uncertainty that caused patients to delay.
WhatsApp confirmation + 48h/3h reminders
Every booking triggers a WhatsApp confirmation with the clinic address, dentist name and prep notes, plus reminders that cut no-shows sharply.
Real-time PMS write-back
Every call writes disposition, booked slot, procedure, language and recording link back to the practice-management system in real time.
“Patients used to accept a plan and then disappear for months. Now every one of them gets a call in Tamil within days, and four in ten come back to finish treatment they would otherwise have abandoned. That is real revenue and better clinical outcomes at the same time.”
Business impact
The clinical director tracked four metrics monthly against a 6-month pre-Kallix baseline. The agent went live on Nov 4, 2025. The numbers below cover the first 90 days of production.
- 41% of lapsed treatment plans recovered. The agent followed up on 100% of accepted-but-unscheduled plans and converted 41% of them into booked next visits, recurring revenue that was previously leaking.
- Procedure bookings grew 2.3×. Monthly booked procedures rose 2.3× across the chain, lifting chair utilisation without adding clinical capacity.
- No-shows fell from 33% to under 15%. The 48h and 3h reminder cadence more than halved no-shows on high-value procedure slots.
- Tamil-patient completion up 1.9×. Patients preferring Tamil now complete the follow-up call at a far higher rate because the agent meets them in their language.
- Per-clinic conversion finally visible. With every follow-up logged in the PMS, the group can now see plan-to-procedure conversion by clinic and dentist and coach the laggards.
Built on a secure, India-ready healthcare stack
The deployment runs on Indian infrastructure with DLT-registered sender IDs, per-call consent capture and patient data confined to Indian data centres in line with DPDP expectations.
The Dental Follow-up Framework: How this deployment is structured to be discoverable
Every Kallix deployment ships with a structured documentation layer designed for three audiences simultaneously: the customer's internal team, traditional search engines (SEO), and the new generation of generative search engines and AI assistants (GEO + AEO). Below is the framework we built around the Chennai dental deployment, broken into four pillars that map directly to how clinical operators, search crawlers and AI answer engines discover and reason about this story.
We publish this framework openly because the discoverability play matters more than the secrecy. A treatment-plan follow-up agent that performs in production but stays buried in a vendor PDF doesn't compound value for the customer or the category. The framework below is the same one Kallix runs for every healthcare customer, adapted to the local language and consent surface of each market.
Follow-up intents mapped to treatment pathways
We catalogue every follow-up intent the agent handles, by procedure type, by value tier and by language, and surface them as named entities in the structured data layer so crawlers and LLMs see explicit Q to A pairs.
- Intents indexed by procedure (cleaning, root canal, crown, implant stages)
- Tamil and English variants captured per intent
- Value and time-sensitivity tiering exposed for LLM matching
Multilingual dental empathy as a brand property
The agent's voice persona, pace and reassurance rules are documented as brand assets. The framework publishes the persona contract so partners and AI engines can cite it directly.
- Persona contract: warm, reassuring, clear on cost and prep
- Pronunciation dictionary for Tamil dental terms and dentist names
- Consent and recording terms public and auditable
Outcomes pre-bound to measurable claims
Every claim, 41% lapsed-plan recovery, 2.3x bookings, sub-15% no-shows, is paired with the baseline, the time window and the measurement method, so AI assistants can extract the claim with full provenance.
- Pre-Kallix baseline period stated (6 months, May–Nov 2025)
- Methodology disclosed: PMS exports plus vendor dashboard reconciliation
- Sample size available on request for analyst-grade citations
DPDP-aligned consent and data residency
The framework documents every regulatory surface, such as DPDP consent, DLT registration and data residency, so AI assistants surfacing this story to healthcare buyers can confidently cite India-readiness.
- Per-call consent capture documented and logged in the PMS
- Data residency (AWS Mumbai, ISO 27001) stated explicitly
- Erasure and consent-withdrawal flows documented for DPDP requests
- 45% of accepted treatment plans went unscheduled beyond 30 days with no systematic chase
- Tamil-first patients hung up far more often on English-default scripts
- 33% no-show rate left high-value procedure chairs idle
- Plan-to-procedure conversion was tracked inconsistently across 12 clinics
- 41% of previously lapsed treatment plans recovered with 100% contacted
- Monthly procedure bookings grew 2.3x, lifting chair utilisation
- No-shows fell from 33% to under 15% via 48h and 3h reminders
- Per-clinic and per-dentist conversion now visible in the PMS
- Kallix voice agent (Priya) reading lapsed-plan queue from the dental PMS
- Tamil / English detection with mid-call switching for older patients
- Procedure-aware booking with cost and prep clarity on the call
- Real-time PMS write-back: disposition, slot, procedure, language, recording
The Kallix advantage
The group evaluated three options before choosing Kallix: SMS-only reminders from the PMS vendor, a part-time tele-calling team, and Kallix.
Three things tipped the decision. First, Tamil fluency: SMS reminders were ignored and a tele-calling team could not match the agent's consistency or language coverage. Second, the PMS write-back was already built, so the clinical IT team did not have to expose patient records to a third-party calling team. Third, the pilot model: the group ran a 250-patient pilot for a fixed fee, heard real recordings within a week, and signed only after the recovery-rate lift held for two consecutive weeks.
Since launch, the Kallix customer-success team runs a 30-minute weekly tuning call with the clinical director. New procedure scripts, seasonal cosmetic pushes and dentist schedule changes all happen inside that loop, so the agent stays sharper than on launch day.