How a Hyderabad multi-specialty clinic chain hit 92% patient-recall confirmation with AI voice agents

The network runs 8 multi-specialty clinics across Hyderabad and Secunderabad, covering general medicine, paediatrics, orthopaedics, dermatology and diagnostics, with roughly 140 staff and a combined patient base of over 90,000 active records.

Revenue depends on chair utilisation and recurring follow-ups. A diabetic patient who skips a quarterly review, a post-op orthopaedic case who misses a dressing change, a child overdue for an immunisation booster, each missed recall is both a clinical risk and lost revenue. The front desk was running recall manually from a spreadsheet, calling patients between walk-ins.

In late 2025 the operations head calculated that the two receptionists handling recall could realistically reach only 40% of the due list each week, and that no-shows were costing the group an estimated 18 lakh rupees a month in idle consultant time. They wanted an always-on layer that could call every due patient in their preferred language, confirm or reschedule, and write the outcome straight into the HIS.

The manual recall model had four compounding failure modes: incomplete coverage, language friction, no-show leakage and no clean data trail for clinical governance.

Kallix deployed an AI voice agent named Lakshmi, a warm Telugu-Hindi-English voice, that pulls the daily due-recall list from the HIS, calls each patient in their preferred language, confirms or reschedules, and writes every outcome back in real time. The full build, from discovery to production, took 16 working days.

Operations tracked four metrics monthly against a 6-month pre-Kallix baseline. The agent went live on Oct 14, 2025. The numbers below cover the first 90 days of production.

The deployment runs on Indian infrastructure with DLT-registered sender IDs, consent capture on every call, and patient data confined to Indian data centres in line with DPDP expectations.

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 Hyderabad clinic 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 patient-recall 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.

The group evaluated three options before choosing Kallix: a generic IVR reminder add-on from their HIS vendor, an offshore call-centre BPO, and Kallix.

Three things tipped the decision. First, Telugu fluency: the IVR add-on offered only flat text-to-speech that older patients found robotic, while Kallix's Telugu voice and mid-call switching kept them engaged. Second, the HIS write-back was already built and tested, so the clinical IT team did not have to write integration code or expose patient data to a third-party BPO. Third, the pilot model: the group ran a 300-patient pilot for a fixed fee, heard real recordings within a week, and only signed the production contract after the confirmation-rate lift held for two consecutive weeks.

Since launch, the Kallix customer-success team runs a 30-minute weekly tuning call with the operations head. New specialty scripts, seasonal immunisation pushes and consultant schedule changes all happen inside that loop, so the agent stays sharper than on launch day.