AI-powered CRM and operations platform for enterprises.
From order to dispatch —
Nazorat (Uzbek for oversight, control) is the first and foundational NEUROUZ product. It’s a managing AI system built around a team of specialized AI agents. It unifies sales, operations and control in one loop — what used to rely on people and spreadsheets. The consumer product Hammasi AI was later built on the same Nazorat architecture (ASCO_BRAIN.md).
This is how it runs in the real deployment at ASCO Holding. A person initiates the order — Nazorat AI takes over the rest: execution, access and control.
The client submits a request: choosing the warehouse and volume.
Nazorat AI automatically records the client and the request, and opens the deal.
A truck arrives at the warehouse — a smart camera reads the plate and matches it against the request.
On a match, the barrier opens automatically — with no human in the loop.
Dispatch begins; AI video monitoring records and controls the process in real time.
The system compiles a report; management receives weekly and monthly operations summaries.
Capturing requests and leads, logging them, running the deal and follow-ups — so deals never stall.
Plate recognition, automatic access control (barrier), and dispatch control on site.
Smart cameras monitor on-site processes in real time and help catch deviations.
Automatic weekly and monthly reports — full operational transparency for management.
Nine specialized agents work around the central NAZORAT core, each in its own area of responsibility. Together they form a single operational system for the enterprise.
Vehicle plate recognition and dispatch video analytics across the site.
ASCO_BRAIN.md (v2.0, 13 sections). The same BRAIN.md approach is reused in Hammasi AI.
AIRNET.uz (Tashkent) — data residency in Uzbekistan, on servers close to the client.
FastAPI backend, agent reasoning on Claude API, data in PostgreSQL, notifications via Telegram.
We deployed Nazorat AI at ASCO Holding — a major Uzbek coal holding. The system runs in production on the client’s servers and drives the full cycle: from the request and gate plate-recognition to dispatch control and reports for management.
In practice it helps catch what used to slip through and run operations more tightly — sales, access and dispatch all sit in one transparent loop.
Proved the AI-agent architecture on a real enterprise client.
Scales the same approach to a mass audience via Telegram.
We’ll show how the system covers sales, access, dispatch and reporting in one loop.