Expensive support operations
80% of call-centre operators handle the same routine questions day after day. People burn out, make mistakes at peak hours, and support costs keep growing.
Resident of IT Park · Tashkent
AI for companies building the future today.
An AI engineering company. We design, build and deploy intelligent systems for banks, retail, TravelTech and the public sector — from audit to production launch.
Philosophy
"
We don't sell demos. We build systems that become part of the customer's business — and stay under their control.
Dilmurad Tillakhodzhaev · CEO · Ilbirs AI Lab
Key figures
/01
12+
Senior AI Engineers
A team with hands-on experience building high-load systems in fintech and enterprise.
/02
100%
In-house development
No off-the-shelf AI platforms. Architecture, code and data stay with you.
/03
24/7
AI automation
Agents handle routine conversations and reduce cost per acquisition.
The problem
Why AI isn't delivering the expected return
Most AI projects stall at the pilot stage. Usually because of three systemic issues.
Unreliable model answers
An uncontrolled LLM answers confidently but inaccurately: invents prices, confuses promotions, promises customers things your system doesn't actually offer.
Disconnected integrations
CRM, ERP, hotel PMS, payment gateways and internal databases aren't ready to work in a single synchronised loop in real time.
Expertise
Enterprise-grade AI architectures
Designed to run in demanding technical environments.
Conversational AI
Next-generation LLMs in customer channels and internal knowledge bases: WhatsApp, Telegram, web chat, voice.
- Multimodal context
- Fine-tuning for your domain
Backend integration
AI modules embedded directly into your core systems. Fault-tolerant, zero-downtime, live-traffic migration.
- Zero-downtime migration
- API-first architecture
Predictive analytics
Demand forecasting, anomaly detection and decision-support models — for planning, risk and operations.
- Real-time anomaly detection
- Bayesian forecasting
Flagship product
A production-ready, in-house AI platform for the hospitality industry
Rethinking the travel market through algorithmic trust and seamless digital logistics. Manzilla is our path to global scale.
98%Valuation accuracy
4.2xFaster deal closing
Industries
Industries we work in
Our AI systems are embedded in mission-critical processes across sectors — solving concrete operational tasks.
Banking & Finance
Automated fraud detection and credit scoring with transparent decision logic.
Retail
Inventory optimisation and personal recommendations based on real customer behaviour.
Logistics
Route optimisation and supply-chain resilience modelling under seasonal peaks.
Health Tech
Medical imaging analysis and clinic resource planning based on patient flow.
Manufacturing
Predictive equipment maintenance and computer-vision-based quality control on the line.
Gov Tech
Public-service automation and decision-support models for city and government operations.
Process
How we work
01
Audit & Discovery
Technical review of existing systems, data and processes.
02
Architecture design
A tailored AI-solution blueprint sized to your scale and stack.
03
Pilot integration
Controlled-environment testing on real datasets.
04
Launch & support
Full production rollout with monitoring and 24/7 support.
Team
Team
Dilmurad Tillakhodzhaev
CEO
A strategist and executive with 25+ years of experience: leadership roles at The Coca-Cola Company (CIS), CEO of the SOF group of companies, consulting on investment and corporate governance in Central Asia.
Baxtiyor Yunusov
Deputy CEO
13+ years of legal and analytical work: operations, leading departments, negotiations. An orientalist; languages — Uzbek, Russian, Dari.
Fedor Tikhonov
CTO
CTO of MBusiness (MBank). 13+ years leading development: built the mobile banking of Optima Bank and CenterCredit Bank, led mobile development at Kolesa.kz / Krisha.kz.
Oleg Bavykin
Mobile Developer
Faculty of Mechanics and Mathematics, NSU. Maintains apps on Android and iOS and builds new ones with Flutter.
Murat Kurbanov
Product Manager
Product manager and business analyst with a medical background: digitising healthcare in Kyrgyzstan (HIS, LIS, Sanarip Clinic), fintech at Optima Bank, the mashina.kg auto-lending product integrated with MBank.
FAQ
Frequently asked questions
How long does a standard AI audit take?
Usually 1–2 weeks. In that time we review existing systems, data and processes, then draft the implementation roadmap.
Do you offer on-premise deployment?
Yes. We deploy both in the cloud and in the customer's closed perimeter (on-premise) — with full control over data and models.
Can your systems integrate with SAP or Oracle?
Yes. We use API-first architecture and connectors to enterprise systems (SAP, Oracle, 1C, local CRM/ERP/PMS).
Start a project
Let's talk about your AI project.
Tell us about the task — we'll show which AI approach fits your situation and what can be shipped within a quarter.
Start a project