The company and the task
Client: A national residential and commercial property brand: more than 50 developments across 10 regions and around 5,000 customer enquiries a month. The name is under NDA.
The company is not named: the project is under NDA.
Industry: Construction · Year: 2025
Customers waited too long for a shortlist: a manager spent an average of 20 minutes on one enquiry, and a new employee up to 30 minutes looking for current information. The first shortlist had to be faster, the recommendations more accurate and the managers less loaded.
Before
Information about the developments was scattered; managers looked up price, floor area, location and availability by hand and got recommendations wrong.
What we did
- Data on 50+ developments was structured: price, floor area, location, availability.
- The bot was trained to understand enquiries like “a 2-bedroom flat in the centre under 12 million”; it was integrated with the knowledge base, the CRM and the messenger — it works on the website and in Telegram.
- A pilot with 5 managers, adjustments to the answer logic, and a briefing for the team on checking the difficult cases.
Results in numbers
What the client says
«The quality of customer service has clearly improved, and fewer mistakes in the sales team’s work improved the company’s figures. Customers note how relevant the properties we offer are, even at that speed of reply.»
