Cybersecurity in on-premises environments presents a growing challenge for organizations lacking specialized digital protection teams. The complexity of current threats demands intelligent solutions that can guide administrators without requiring deep security expertise. In this context, knowledge-based decision support systems (DSS) emerge as key tools. This article proposes an original approach: a knowledge-based multi-agent system for security recommendation, designed to balance control coverage with resource efficiency, avoiding both under- and over-provisioning.
The system is modeled as a non-cooperative, non-zero-sum game, grounded in a Multi-Agent Influence Diagram (MAID). Each agent represents a security dimension — such as access control, data protection, monitoring, etc. — and collaborates to find the optimal combination of control sub-families that meet user requirements. Using no-regret online learning, the system explores the decision space and quickly converges to solutions that maximize coverage satisfaction. Experimental results show that it is possible to achieve 99% coverage using only 65% of software-implementable controls, with response times between 1.2 and 35.7 seconds; while using 29% of controls yields between 73% and 77% coverage in 0.8 to 13.8 seconds.
This approach goes beyond traditional rule-based recommendation. By integrating multiple intelligent agents, the system dynamically adapts to the changing needs of the organization. Each agent carries expert knowledge extracted from consolidated sources, both academic and from the information security (InfoSec) industry. The unification of this data into a curated dataset allows the DSS to offer precise recommendations without complex configurations. For companies looking to strengthen their security posture without duplicating investments, this technology represents a significant advance.
From a technical perspective, implementing this multi-agent system requires a robust and scalable infrastructure. This is where the expertise of Q2BSTUDIO as a software and technology development company becomes essential. The company offers custom software that integrates artificial intelligence and cybersecurity modules, deployable both on-premises and in the cloud. For instance, using cloud services like AWS or Azure allows elastic scaling of agents, processing large volumes of security data in real time. Additionally, integration with Business Intelligence tools such as Power BI facilitates visualization of coverage metrics and continuous monitoring of control compliance.
The convergence of AI and cybersecurity is no coincidence. Intelligent agents can learn from past incidents and proactively adjust their recommendations. Instead of applying a static catalog of controls, the system proposes sub-families that adapt to each company's risk profile. This is especially valuable for SMEs that lack a dedicated security team but need to comply with regulations like GDPR or ISO 27001. Agent-based recommendation minimizes human error and reduces the time needed to design the security architecture.
In terms of performance, the obtained results demonstrate the practical feasibility of the system. With datasets of varying sizes, the no-regret algorithm maintains stable convergence, even when new requirements are introduced. The simultaneous nature of the game ensures that each agent negotiates its own coverage without conflicts, achieving a Nash equilibrium that satisfies the end user. This model contrasts with sequential approaches, which often generate suboptimal or redundant solutions.
For organizations wishing to implement a similar solution, Q2BSTUDIO offers consulting and development services in artificial intelligence applied to cybersecurity. The company combines its expertise in cloud AWS/Azure, Business Intelligence, and process automation to build custom systems that maximize protection with minimal resource usage. For example, integrating Power BI allows security managers to monitor in real time the effectiveness of recommended controls, facilitating strategic decision-making.
In conclusion, the knowledge-based multi-agent system for security recommendation represents a qualitative leap in enterprise cybersecurity management. By leveraging multi-agent game techniques and online learning, it achieves an optimal balance between coverage and efficiency. Companies like Q2BSTUDIO are leading this transformation, offering personalized solutions that integrate AI, cloud, and BI in a unified framework. Security is no longer a luxury reserved for large corporations: with intelligent and affordable tools, any organization can effectively protect itself against digital threats.
