★ Our strength

1C database hygiene: counterparty deduplication and completing item data

Over time a 1C database gets cluttered: duplicate counterparties, records without a BIN or bank account, empty attributes and an unfilled product identifier for ESF. The assistant finds the problems, prepares draft fixes and brings order, but writes to the database only after your confirmation.

Discuss your task WhatsApp

What we found in a real database (2248 counterparties, 13 186 items)

Counterparties: 44 groups of duplicates by BIN plus 17 placeholder records without a BIN; 79 without a BIN or IIN; 702 (31%) without a main bank account.
The product identifier for ESF is empty for all 13 186 items: the biggest gap for ESF.
Attributes are almost empty: dimensions for 18%, weight for 12%, resolution and IR illumination for 5%, PoE for 1.4%; Kazakh names are empty for almost all items.
There are inconsistencies: a Ubiquiti record with the brand “Ruijie” and someone else's website name, three products in one record because of copy-paste.

What it does

Finds duplicate counterparties by BIN and placeholder records without details (candidate report → confirmation → merge).
Completes records: pulls the bank account and BIN from linked documents.
Fills in item attributes: AI prepares a draft from the name and datasheet.
Selects the product identifier for ESF by the HS code (TN VED) from the government catalog.
Prepares Kazakh item names (AI translation and transliteration).
Audits record consistency: catches mismatches of brand, name and content.
Works as hygiene at the input, not a one-time cleanup: the database stays clean afterwards.

How it works4 steps

Database diagnostics

The assistant goes through counterparties and items and compiles a report: duplicates by BIN, records without details, empty attributes, an unfilled identifier for ESF.
Merge candidates

A candidate report is generated for duplicates, for example 44 groups by BIN. You confirm which records to merge, and only then does the merge happen.
Filling in attributes

AI prepares draft attributes from the name and datasheet, selects the ESF identifier by the HS code (TN VED) from the government catalog and proposes Kazakh names.
Confirmation and writing

Each fix arrives as a draft for confirmation. Only what you approve is written to the database: no automatic edits.

Result

44 groups of duplicate counterparties by BIN found in a real database

100% → we fill it in the ESF identifier was empty for all 13 186 items

from 1–18% to complete item attributes are filled in

Who it is for

For companies whose database is cluttered with duplicates and records without details.
For those preparing items for ESF who are stuck on an empty product identifier.
For businesses that need a clean database both for the website and for quick selection by an engineer.

Frequently asked questions3

Will the assistant merge or delete records by itself?

No. First you receive a report of merge candidates, then you confirm what to merge. Merging and any change happen only after your “OK”: nothing is deleted automatically.

Where does the product identifier for ESF come from?

The assistant selects it by the HS code (TN VED) from the government catalog and proposes it as a draft. Right now this field is empty for all items: we fill it in for your review, not at random.

Is it a one-time cleanup or ongoing?

The principle is hygiene at the input, not a one-time cleanup. The assistant helps both to bring order now and to keep the database clean later, so duplicates and empty attributes do not pile up again.