Jason Oehley, regional lead at cyber security firm Arctic Wolf. (Image: Supplied) International cyber security leader Arctic Wolf will showcase its agentic security operations centre (SOC) model at the forthcoming annual ITWeb Security Summit 2026 in Johannesburg – technology that is designed to balance human and artificial intelligence (AI) to leverage trustworthy data, train reliable agents, effectively manage automation and counter the high failure rate of AI initiatives.Speaking ahead of the summit, Jason Oehley, regional lead at Arctic Wolf, said many AI initiatives fail because of a lack of data, reliable AI and projects that are based on synthetic models. “This means the initiatives are not designed to address real world risk,” he added. “Organisations require huge amounts of data, teams to train AI and real-world scenarios to build trust. Without this, AI will act on its own and if left to do so, increases the risk of hallucinations and of erroneous data.”Traditional SOCs are not able to match the pace and precision of modern threat actors, especially those that leverage AI to accelerate and scale attacks.These SOCs are also constrained by tiered teams, alert overload, fragmented tools and linear workflows.These limitations make AI-powered threats even more dangerous as threat actors are able to leverage the technology to keep ahead of defences.Security teams also need AI, but current solutions are often hard to trust and deploy. Hallucinations, weak reasoning and unreliable results reduce confidence, while data preparation, workflow adaptation, validation and ongoing governance add strain to already stretched teams.This is why Arctic Wolf advocates the agentic SOC model.The company said unlike the traditional SOC, in which humans do the work, an agentic SOC is a security operations model in which AI agents execute the primary workflows of detection, investigation and response.An agentic SOC is defined by AI-led execution, continuous operation, parallel investigation, evidence-based reasoning, environment-aware decision-making and governed autonomy.According to Arctic Wolf, this is not a feature added to the traditional SOC, it is a different operating model in which AI becomes the primary execution layer and humans provide direction, control and oversight.Oehley explained that the company has an authoritative and oversight AI model, which ensures the correct agents are involved – either AI or human. “The judge is to validate the outcome. Should there be additional confirmation needed, the judge would route this to a human for validation. This helps ensure a trusted AI model,” he added.Oehley said many cyber security solution vendors claim to have released AI-powered defence products, but the AI is built on top of existing solutions and not fully integrated from the beginning.Arctic Wolf plans to fully introduce its agentic SOC model at the summit and explain to delegates why this represents competitive advantage in a digital, AI-driven market.“This is an ideal opportunity for companies to engage with us and find out more about technology that is available to them to strengthen their defences… we are positioned to provide valuable insight,” added Oehley.For more information and to register, visit ITWeb Security Summit 2026.About ITWeb Security Summit 2026ITWeb Security Summit 2026 was held at Century City Conference Centre, Cape Town on 26 May 2026 and is being held at Sandton Convention Centre in Sandton, Johannesburg from 2-4 June 2026.Themed: ‘Redefining security in the face of AI-driven attacks, fragile supply chains and a global skills gap’, the 21st annual edition of ITWeb Security Summit will continue in its tradition of bringing leading international and local industry experts, analysts and end-users together to delve into the specific threats and opportunities facing African CISOs, security specialists, GRC professionals and anyone else who is responsible for securing their organisation from cyber attacks.Register today. Visit here for Johannesburg.