The company and the task
Client: 21 medical centres in 13 cities of Russia, more than 3,000 staff, part of the Mother and Child group.
Industry: Healthcare · Year: 2026
The network’s contact centre was a black box: the sample of calls listened to was about 1%, and in some periods 0.04%. The score depended on the auditor’s mood, and outsourcers did not know the particulars of the clinics. Scoring one call by hand cost 65 roubles, and checking all the traffic would have taken 52 extra staff.
Before
Reasons for cancellation, objections and unmet demand were collected by hand and were imprecise. “We knew we were losing patients, but we did not know where,” recalls the operations head of the contact centres.
What we did
- Automatic scoring of 100% of calls against a single checklist of more than 30 points; more than 30 dashboards by city, agent and period.
- Analysis of conversion on in-scope calls, of the root causes when it fails (need never established, no qualifying questions, objections left unhandled) and of unmet demand.
- Analysis of cancellations: whether the agent offered to reschedule and whether the patient was retained is recorded. Real-time prompts for agents, integration with the telephony.
- Tags first (2024), then language models for semantic scoring (2025). By October 2025 there were 11 cities on the platform.
Timeline: 2024–2025: launched within a year, scaled to 11 cities
Results in numbers
What the client says
«Speech analytics gave us 100% coverage, objectivity and transparency in scoring. We closed an entire department of specialists, cut the cost from 65 to 2.6 roubles per call, and started growing where we used to be guessing. It was not easy, but the result exceeded our expectations.»
