Skip to content

Applied AI lab

This is where we build our own products. What we learn goes into our client projects.

~/productos

Our products

What we build for ourselves, and then bring to client projects.

$ ls productos/

siteflow/ legalos/

$ cat siteflow/README

SiteFlow

In development

Turns construction reports sent over WhatsApp into logs, indicators, alerts and automatic reports.

Problem: On construction sites, reports arrive as loose photos and messages nobody has time to consolidate.

$ cat legalos/README

LegalOS

In development

Platform for lawyers and law firms that brings together clients, cases, documents and AI assistants, built for Colombia first.

Problem: Case and client information is scattered across folders, inboxes and spreadsheets.

$

pqrs-flujo.log

How AI works in a real process

This is how every step of an AI-assisted citizen request gets logged. A person always makes the final call.

[09:14:02]PORTALA citizen request comes in

[09:14:03]AIClassifies the request type and the responsible area

[09:14:05]AIDrafts a reply within the legal deadlines

[09:21:40]STAFFReviews, edits and approves

[09:21:41]SYSTEMSends the reply and stores the audit trail

ia-responsable.yml

Responsible AI

# Data under control: Where data is processed and stored is defined in writing before we start.

data_under_control: defined_in_writing

# Human approval: Decisions that affect people are approved by someone on your team.

human_approval: always

# Traceability: Every AI suggestion and every decision is logged.

traceability: complete

Want to pilot a product in your operation?

$ limitless pilot --request

Talk to us