Kotib Analytics: AI-powered Call Quality Evaluation for Assalom Estate
- Customer
- Assalom Estate
- Project manager on the customer side
-
- IT Provider
- Kotib.AI
- Year of project completion
- 2025
- Project timeline
- September, 2025 - October, 2025
- Project scope
- 50 man-hours
- Goals
-
The project aimed to improve customer experience and sales team efficiency through automated AI-based call evaluation. Before implementation, Assalom Estate managers manually listened to calls to assess quality, which was time-consuming and inconsistent. Kotib Analytics introduced AI-driven speech recognition and quality scoring, enabling management to track key metrics such as greeting standards, empathy, product knowledge, and conversion cues in real time.
- Project Results
-
Results:
– Over 85% of calls are now automatically evaluated by AI.
– Average call quality score increased by 22%.
– Supervisors’ time spent on call audits reduced by 70%.
– Real-time insights helped identify weak points in customer interactions and improve conversion rates.The solution demonstrated how AI can enhance business transparency and customer focus in the real estate industry, setting a benchmark for other local companies.
The uniqueness of the project
The uniqueness of this project lies in adapting an AI speech analysis system for the real estate industry in Uzbekistan — a market where digitalization of customer communication is still in its early stages. The AI model was trained to understand local accents and mixed Uzbek-Russian speech, ensuring accurate sentiment and compliance analysis. Another distinctive feature was integrating business metrics directly into dashboards, linking communication quality with lead conversion results.- Used software
-
Kotib Analytics (AI-based call analysis platform), Google Cloud Speech-to-Text, PostgreSQL, REST API integration with internal IP-telephony.
- Difficulty of implementation
- The main challenges included speech recognition accuracy for multilingual communication (Uzbek and Russian mix) and limited labeled data for training. The team solved this by collecting real call samples and continuously fine-tuning models. Integration with the company’s telephony required custom API work. Despite these challenges, the full deployment was completed within 3 months.
- Project Description
-
The project involved implementing Kotib Analytics — an AI-powered system for automatic evaluation of sales calls — in Assalom Estate, one of the leading property developers in Uzbekistan.
Implementation steps:
– Connected Assalom Estate’s call recording system to Kotib Analytics.
– Customized AI models to recognize mixed Uzbek-Russian speech and real estate-specific vocabulary.
– Designed quality criteria aligned with company standards: greeting, tone, needs identification, property matching, and closing.
– Created dashboards for managers to view individual and team performance.
– Conducted staff training and pilot evaluation of 1000+ calls.Results:
– Over 85% of calls are now automatically evaluated by AI.
– Average call quality score increased by 22%.
– Supervisors’ time spent on call audits reduced by 70%.
– Real-time insights helped identify weak points in customer interactions and improve conversion rates.The solution demonstrated how AI can enhance business transparency and customer focus in the real estate industry, setting a benchmark for other local companies.
- Project geography
-
Uzbekistan (Tashkent)