Andre Troskie, EMEA field CISO at Veeam.
Agentic AI is both a powerful opportunity and a massive risk, requiring completely new approaches to how organisations approach data resilience and trusted AI.
This is according to speakers at an executive roundtable on trusted AI and data resilience, hosted in Johannesburg by Veeam in partnership with ITWeb.
Nerushka Bowan, technology and privacy lawyer and founder of the LITT Institute, highlighted how AI-enabled threats had extended their reach across both the virtual and the physical realm, with tools such as AI glasses being increasingly misused. “It’s hard to keep up with change, especially when it comes to AI,” she said. “In the wrong hands, it could be catastrophic.”
Andre Troskie, EMEA field CISO at Veeam, noted that the World Economic Forum’s Global Cybersecurity Outlook 2026 expects AI to be the most significant driver of change in cyber security in the year ahead, with AI vulnerabilities accelerating at an unprecedented pace.
The WEF Global Risks Report 2026 also puts cyber insecurity among the top 10 biggest short-term risks, with ‘Adverse outcomes of AI’ the risk with the largest rise in ranking over time, moving from number 30 on the two-year outlook to number five on the 10-year outlook.
“Agentic AI is by far the biggest opportunity, but also the biggest threat, ever to emerge in the world of IT. It creates a world of increased risk, and things are ramping up exponentially,” he said.
As organisations enter the agentic AI era, success will depend on ensuring that data is accurate, secure, governed and always available, he said.
“AI is growing faster than its guardrails, and organisations now need the ability to manage and recover as fast as machine-speed attacks. What we are seeing is a converging data crisis and visibility gap, and when we apply AI at speed, the gap grows wider. Yesterday’s security systems and guardrails will not protect us from tomorrow, and organisations across the world are not ready for what is coming.”
Troskie said: “Organisations are not where they need to be, if you consider the speed and scale of AI and the resilience challenges that come with that. And it’s not just Africa – it’s the world. We are all struggling with the same thing: the control objectives are very well understood, but the controls that we need to implement today and tomorrow are going to be very different from what they were before. We need to move to a world of continuous, autonomous assurance.”
Troskie showcased Veeam’s new DataAI Command Platform, a unified data and AI trust infrastructure to secure data, govern AI and automate compliance for the agentic AI era. The solution offers one unified platform powered by Veeam DataAI Command Graph for data and AI security, governance, privacy, compliance and resilience. It maps relationships across data, identities, systems and AI to create a unified layer of intelligence.
The Veeam DataAI Command Platform allows organisations to see all their data, backups, identities and AI systems, continuously discover and classify sensitive data, and identify where it is exposed, over-shared or at risk. The platform allows them to prioritise and remediate high-risk data exposures, and even gain visibility into AI systems to identify risky activity, and enforce controls to govern how data is accessed, used and shared.
Troskie said: “The new Veeam DataAI Command Platform, launched on the back of our acquisition of Securiti AI last year, brings together the data on one side and the entitlements on the other. We are now integrating Securiti AI’s data security posture management with Veeam’s world-class, leading resilience solutions, so what we will see over the next few months are integrations and implementations of that technology in a modular way. It is the only solution I have seen in the market with this level of scale and completeness, to start dealing with tomorrow’s data resilience and AI challenges.”
