TILAP-IA

TILAP-IA —Técnica Integral para Laborar y Archivar Proyectos asistidos por la Inteligencia Artificial— solves a concrete problem: an environmental study is a long document, written by several hands and subject to review, where every chapter has an owner, a reviewer and a version that has to stand up to scrutiny later. That work usually ends up spread across loose files nobody manages to reconcile. The platform turns it into a single living case file.

The document is written by several people at once, over the same text and with visible cursors: it is real collaborative editing, not a file people take turns with. Each section carries its version history, its threaded comments, its attachments and its status; approval and rejection happen section by section, with history and batch approval.

It ships with 52 system templates for the study types used in Colombia: 15 for environmental impact studies, 15 for management plans, 12 for baseline studies and 10 for environmental assessment of alternatives, built on the ANLA methodology and the Ministry of Environment guidelines. Templates are written in a strict structured language: if the hierarchy is wrong, the document is rejected rather than silently producing a broken structure.

Field data capture keeps the format whoever requires it expects. Forms are filled in on screen, but the output goes back into the original Excel file with its layout intact —not a generic field-and-value table— and is converted to a faithful PDF. So as not to write one program per form, a language model works out the mapping between each field and its cell just once; that mapping is reviewed and stored, and generating each project is deterministic, without consulting the model again.

The final document consolidates the approved sections and comes out in ODT, DOCX and PDF. Before generating it, the system validates what is missing and which forms are incomplete. The format follows the Manual de Seguimiento Ambiental de Proyectos: justified body text and photographic records at the prescribed size. Every generated version is filed away and downloads on its own or with its annexes.

The templates follow the study types in Colombian environmental regulation, which means the document comes out with the expected structure — not that the system certifies anything before the environmental authority: responsibility for the content lies with whoever signs the study. And what does not exist yet is said plainly — cartography, semantic search across the corpus, a charts dashboard, email notifications and a mobile app are planned and not delivered.