Brazilian companies suffered, on average, over 2,600 cyber attacks weekly in early 2025, a 21% increase from the previous year, according to a survey by Check Point Research. In this scenario, the traditional ‘firefighting’ approach in digital security proves insufficient in the face of the speed and sophistication of threats, many of which are powered by artificial intelligence.
‘The future of digital defense requires us to stop waiting for the attack to happen. The answer lies in offensive protection: to think and act like an attacker to find and fix flaws before they are exploited,’ says Rodolfo Almeida, COO of ViperX, an offensive protection startup of the Dfense Group.
From reaction to anticipation: the attacker’s mindset
Offensive protection involves simulating actions of a real intruder to identify vulnerabilities and block paths before they are used. Techniques like red teaming and adversary emulation allow validating technical and logical loopholes, prioritizing corrections based on real risk.
The philosophy of the ‘ethical hacker’, a professional hired to act in an ethical and controlled manner, is already adopted in countries like the USA and members of the European Union, as well as Brazilian initiatives such as the actions of the Central Bank and the Cyber Guardian exercise. ‘This approach goes beyond a simple annual checklist: it embeds security as a strategic function of the organization,’ explains the executive.
The recently enforced National Cybersecurity Strategy (E-Cyber) reinforces this need, elevating the issue to the state level, with pillars of resilience and cooperation.
Artificial intelligence: accelerating offensive protection
AI should not be seen as an enemy but as an ally in the anticipation process. It connects scattered information, from cloud changes to leaked credentials, to indicate where an attack would likely begin.
For Almeida, technology enhances offensive security on three fronts:
- Mapping and prioritization – continuous scanning of the attack surface and highlighting the most critical vulnerabilities;
- Attack rehearsals – simulation of realistic scenarios to test controls and identify failures quickly;
- Accelerated correction – suggestion and automation of solutions to reduce rework and eliminate opportunities for criminals.
“With AI, we find open doors and manage to close them before they are crossed,” emphasizes the executive.
From one-off action to continuous discipline
Almeida emphasizes that technology alone does not solve the problem. It is necessary to abandon the “one-off project” approach and adopt a continuous discipline of managing exposure to threats (Continuous Threat Exposure Management – CTEM).
The executive recommends that companies prioritize metrics that reflect real impact on the business, such as response time and number of blocked attack paths, and adopt ongoing testing and training programs, in addition to sharing knowledge with partners and ecosystem organizations.
“Digital security is not just about protecting oneself, but about building a more resilient digital future for everyone,” concludes Almeida
