Two managers for 500 dealers.
A window-profile manufacturer: two managers couldn't keep up with requests from 500+ dealers — reconciliations, order status, documents, claims — around the clock.
We built a 24/7 Telegram assistant that covers the five main dealer request types. A dealer writes in natural language; the assistant detects the intent, reads the data from 1C over a read-only SQL connection and answers on the spot: a reconciliation, the current order status, a document generated from 1C data right in the dialogue, or an answer from the knowledge base on claims, certificates and products.
Every request and its outcome is logged in Bitrix24, so managers see the full picture without chasing channels. Reconciliations that used to mean a phone call in business hours now get an answer in Telegram at any time, documents are no longer pulled from 1C by hand, and the two managers handle the complex cases while the assistant closes the routine. Payback: two to three months.
What was in the way
- ✗Two managers physically could not cover 500+ dealers
- ✗Reconciliations and order statuses — by phone, business hours only
- ✗Documents were pulled from 1C by hand
- ✗At night and on weekends dealers waited for morning
What the system does
Reconciliations on demand
A dealer writes in natural language — the assistant detects the intent and reads the data from 1C.
Order statuses
The current order state without calling a manager, any time of day.
Documents
Generates documents from 1C data right in the dialogue.
Claims and certificates
Accepts claims and answers from the knowledge base — certificates, products.
History in the CRM
Requests and outcomes are logged in Bitrix24 — managers see the full picture.
How it fits together
How we roll it out.
Requests and access analysis
We study historical requests, verify CRM and API access, and score the automation potential.
A slice of real traffic
The agent runs on part of the live requests with weekly syncs — real dialogues, not a demo.
Metrics and scale-up
We evaluate volume, response speed and escalation share — then expand the agent to full traffic.
Before → after
The full economics of this case — $2.7K in, ~$14K net effect in year one, a before/after table and a security FAQ.