Dossier · supplementary detail

13 bots. The agents don't do the work — they buy us time to do better work on the ground.

×5Extraction

×1Security & storage

×2Surveillance

×2Analysis

×1Data organizing

×1Data reporting

×1Overseeing

Extraction feeds both the publish-to-database step and health monitoring; both feed analysis, then data organizing, reporting, and the overseeing step that reviews the whole cycle.

Time is key

The data evolves as fast as our metrics do. Continuous extraction means a mispricing is caught the day it appears, not the week someone happens to browse the right listing — the edge in the discount thesis is largely a speed edge.

What replaces the manual screen

Instead of screening opportunities across a dozen apps and channels, everything is kept current on a single platform that sends reports and timely notifications. The measurable effect is not better judgement — it is that the judgement gets applied to a complete field rather than to whatever happened to surface that week.

The maths and the AI

The mathematical models used to turn the data into something digestible and readable are written in scripts. The result cannot change without human intervention.

The data is collected, processed, and compared by a local artificial intelligence. It feeds the database, which feeds the mathematical models.

The agents read the data, the maths, and the news to make sure the system stays relevant. Every day. The more time passes, the more solid it gets.