Contracts, offers and purchase orders drafted by private models trained on the company's own contracts.

SITUATION

A supplier contract campaign meant hundreds of documents assembled by hand from the same templates.

Each year the company contracts its wood suppliers in a campaign: the contract with its delivery conditions, the supplier declaration, the addenda, the price matrix for each payment term. Every document was assembled by hand, and every change in terms meant editing all of them again.

WHAT WAS BUILT

A contract hub that produces the campaign's documents from the company's own models: contracts, exhibits, offers and purchase orders, with the current contract models and delivery conditions, the supplier declaration, the addendum types including fixed-price, and price matrices for each payment term. Private AI models trained on the company's own contracts draft and review the documents, and every one goes to the person who signs it.

HOW IT RUNS

On-premise, inside the company's network, with local authentication and an audit log on every access. No contract, price or supplier record leaves the company's hardware.

WHAT CHANGED

The campaign's documents come out of one system, in the current terms, and a change in terms is made once.

THE STACK

MODELS
Private language models, trained on the company's contracts

DOCUMENTS
Contracts, exhibits, offers, purchase orders, addenda

ACCESS
Local authentication, role-based, audit log

RUNS ON
On-premise, inside the company's network