Nowadays, the anomalies found in the prevention of fraud are aimed at detection tools that exist in themarket of each one of the financial entities; many of them are focused onalgorithms predisposed for their use and implementation. However, some of them do not have an effective learning in the face of the needs of the phishing phenomenon as a crucial element in the face of the challenges of these sites in Colombia, this is a reason that does not allow their parameterization and and then prosecute criminal behaviour. The adaptive algorithm model allows detecting and investigating phishing from a tactical orientation that allows identifying these characteristic patterns.

Keywords : Machine Learning, model, social engineering, artificial intelligence,cyber-research, OSINT.