AI adoption is accelerating, but initiatives without proper governance, visibility and trust risk exposing organisations to significant disruption.
This was a recurring theme at an executive roundtable held alongside the ITWeb Security Summit 2026 in Cape Town. The event, hosted in partnership with Veeam Software, took place at Bistro Sixteen82 at Steenberg Estate.
Setting the scene, Andre Troskie, EMEA Field CISO at Veeam, explained that a cyber storm is building. Rising data volumes, increasing complexity, sophisticated cyber crime, persistent skills gaps, identity sprawl and limited oversight are all placing pressure on organisations to strengthen resilience.
“In Veeam’s view, this data crisis is being driven by three major gaps. A visibility gap means organisations lack a clear understanding of what is happening with their data. The AI trust gap highlights the need to establish control over AI environments. This in turn creates a resilience gap, raising a critical question: can you recover quickly and at scale if something goes wrong?”
To close these gaps, organisations must implement controls that are aligned to risk and designed to be both effective and efficient. These controls must also support core regulatory requirements, including accountability, incident response, risk management and supply chain risk.
Navigating visibility, trust and resilience gaps
The room then split into smaller groups to discuss topics including what minimum viable data visibility looks like and how to enforce zero trust when AI pipelines and automated workflows require access to critical systems. Participants also explored the challenge of recovering data without re-introducing compromised data, and whether they have the ability to verify that restored data is clean.
Attendees agreed that the pace of change and exponential data growth make it increasingly difficult to keep up, particularly when implementing the policies needed to protect the business. While AI is helping to drive efficiency in some areas, it is also introducing new risks around data management and integrity.
Understanding data flows and maintaining visibility over how data moves in and out of the organisation remains a common challenge. Participants noted that where full monitoring is not possible, organisations must prioritise the data sets critical to operations and apply appropriate risk management controls. Without this, security teams risk being seen as barriers to innovation. Education therefore plays a key role in enabling responsible AI use.
Approaches vary. Some organisations are restricting AI use entirely before approving specific use cases, while others are implementing guardrails and managing risk as adoption progresses. In both scenarios, there is clear tension between enabling the business to innovate and ensuring the organisation remains protected.
What is clear from these discussions, said Troskie, is that the challenge is shared. “It is tough, and it is likely to become more complex. But it is also clear that organisations are not facing this alone.”
Closing the session, Ian Engelbrecht, Veeam’s Field CTO for EMEA, demonstrated how Veeam supports organisations by creating what he described as a central nervous system for data. By enabling data discovery and inventory, Veeam helps build a knowledge graph that maps data lineage and provenance across the business.
“The goal is to establish a trust layer between your infrastructure and your future AI environment so you can understand, secure and validate your data before putting it to use.”
Share
