Tahir Latif, data trust and governance lead, EMEA East, Veeam.
As organisations across Africa race to harness AI, trusted and well-managed data has become a critical foundation for success. However, a recent survey has found that while many organisations are confident in their data and recovery capabilities, fewer can consistently demonstrate the visibility, governance and resilience needed to support AI adoption at scale.
This emerged during a webinar on data resilience hosted by Veeam, in partnership with ITWeb. Tahir Latif, Data Trust and Governance Lead, EMEA East at Veeam, and Mark Govender, Manager of Sales Engineering for Africa at Veeam, shared findings from a recent Veeam-ITWeb survey on data confidence and AI readiness, highlighting a significant gap between confidence and operational readiness.
Latif said: “This pan-African survey gathered responses from senior IT, data and cyber security leaders in 14 African countries, largely from financial services, IT and public sector organisations, as well as the automotive, healthcare, energy and utilities, and retail sectors.”
The survey found that 85% of respondents were at least moderately confident they could recover clean, usable data from a cyber incident within agreed business recovery targets.
“There is a difference between believing we can recover and demonstrating we can recover,” he said. “The question is how organisations can substantiate their recovery confidence. The survey found that 32.4% had never performed a full technical recovery test, had not tested in over a year, or did not know when testing last occurred. Meanwhile, 45% of organisations that had completed an end-to-end recovery test reported only partial success, significant gaps or outright failure.”
The findings also revealed a disconnect between confidence in data management and the ability to validate that confidence. While 56% of respondents rated their organisations' ability to discover, classify, govern, protect and retrieve critical and sensitive data as good or excellent, and 50% were very or extremely confident they knew where their most valuable data resided, evidence often fell short of expectations.
Respondents cited legacy systems, data sprawl, shadow IT and incomplete or inconsistent data classification as key challenges. Only 37% could produce an accurate audit trail showing who had accessed their most sensitive datasets across most or all critical systems during the previous seven days.
Govender noted: “Organisations must move beyond traditional backup to data resilience. True resilience is about proving you can recover clean, trusted data and keep critical business services running when it matters most.”
Turning to AI adoption, Latif explained that many organisations are progressing with AI initiatives faster than they are establishing the controls needed to manage them.
“Using AI is AI adoption, not AI transformation,” he said. “Organisations can deploy AI quickly, but that does not automatically mean they have the visibility, governance and resilience needed to use it safely and at scale.”
The survey found a wide range of AI maturity levels, with only 14% of respondents saying AI was embedded into core business processes or autonomous workflows. Just 29% rated themselves as ready to adopt, govern, secure and scale AI initiatives using trusted and well-controlled data.
Mark Govender, manager, system engineering Africa, Veeam Software. (Image: Supplied)
“The challenge is not whether to adopt AI,” Latif added. “The challenge is whether organisations have the visibility, governance and resilience required to trust the data that AI depends on. Maintaining visibility into how AI systems access and use data remains one of the most important governance challenges facing organisations today.”
Govender added: “Organisations need to identify the right AI use cases, establish clear governance controls and ensure trusted data foundations are in place before scaling AI initiatives.”
To help organisations address these challenges, Govender highlighted Veeam capabilities including SureBackup, Veeam Recovery Orchestrator, Securiti AI and Veeam Intelligent ResOps, which combines data intelligence with data protection for Microsoft 365 environments.
He explained: “Veeam’s Data Command Platform delivers full context around data and AI from multiple perspectives. It helps organisations discover and classify data while establishing common policies, classifications, and governance standards across data assets at scale. This provides the visibility and control organisations need to build trusted AI foundations.
“Once organisations understand where sensitive data resides, the next question becomes who can access it and how it is being used,” he said. “Data Exposure Intelligence helps identify potentially exposed systems and prioritise remediation, while Data Access Intelligence provides insight into data entitlements, access patterns and least-privilege access models.”
Latif outlined five key actions organisations can take to bridge the gap between confidence and capability.
“First, organisations must regularly test their recovery capabilities to determine whether clean, usable data and critical business services can be restored within agreed recovery targets,” he said.
“They must also know and control their data by discovering and classifying sensitive information, reviewing permissions and maintaining accurate records of where data resides and who has access to it.
“In addition, organisations should prepare trusted data for AI by understanding what information AI systems can access, whether that data is appropriate for its intended purpose and which privacy, compliance and governance requirements apply.”
He added that successful AI readiness requires clear ownership across recovery, security, data governance and privacy teams, as well as shared accountability for outcomes.
“Finally, organisations should measure progress. Recovery test results, visibility into data usage and the ability to explain, govern and control how data is used all provide important indicators of resilience and AI readiness.”
As AI adoption accelerates, the survey findings suggest that the organisations most likely to succeed will not necessarily be those deploying AI the fastest. Rather, they will be those that can demonstrate trust in the data that powers it through strong governance, clear visibility and proven resilience.