The company and the task
Client: One of the largest property developers in Moscow: every eighth flat sold in the capital, with headcount up fourfold in three years.
Industry: Construction · Year: 2026
Communications were scored by hand: 3% of calls were listened to and 97% went unanalysed. Feedback reached staff long after the event. The previous foreign system left the market, and two pilots with other vendors failed: one could not build reporting from CRM data, the other raised its price during the pilot.
Before
The deciding factor was an in-house LLM module: they needed flexible checklists, analysis of calls, chats and video meetings, and online reporting that drew on CRM data.
What we did
- Process review, a detailed specification, checklists built around a developer’s work; the pilot ran for 8 months.
- Reporting by agent: viewings booked, objections handled, repeat viewings; a dashboard with an agent ranking and their appeals against the automatic scores.
- AI analysis of the reasons people decline a viewing: “too busy”, “thinking about it”, “family circumstances”, “finances” — identified and classified automatically.
- Data for the corporate training system, including an internal quiz; integration with internal systems, analysis of audio, video and chats.
Timeline: An 8-month pilot, scaled up from the start of 2025
Results in numbers
What the client says
«The main goal is growth in sales. The analytics helps us assess how the sales team works, how effective they are at booking viewings, how staff apply what they learned in training and whether anyone is pressuring customers. We identified the reasons customers choose A101 — including some that were not obvious to us.»
