Salvest manages homeowners' associations and receives daily invoices for supplies, elevators, cleaning, maintenance, gardening and other providers.
Each document had to be manually reviewed to identify to which community does it belong, who is the provider, what service does it correspond to, and where should it be filed?.
Salvest Gestiones turns that administrative flow into a virtually automatic process: it receives invoices by email, analyzes the documents, extracts the relevant information, validates the classification and files them in their corresponding location.
Understand
The problem wasn't simply reading an invoice. The same supplier could appear under different names, work for multiple regions, or send documents in different formats. Furthermore, the emails included signatures, images, and other irrelevant attachments, and misclassification could be worse than requesting a manual review.
Choose
A hybrid system was designed: AI interprets the document, The backend uses the database, tax IDs, aliases, known relationships, and business rules to decide how to classify it. When the evidence is insufficient, the document is sent for review instead of making a doubtful decision.
Build
The workflow automates the entire invoice process: email, PDF detection and validation, AI-powered extraction, community, provider, and service identification, classification, and archiving. The system prevents duplicates, learns new valid relationships between communities and providers, and maintains traceability for every decision.
Result
Human intervention is reserved for truly ambiguous cases. The system filters out irrelevant attachments, interprets invoices, identifies suppliers and communities, prevents duplicate processing, and automatically archives documents that can be classified with sufficient security.