CASA Software outlines how fragmented infrastructure and isolated AI initiatives are slowing enterprise technology programmes

Drawing on research from Broadcom and Gartner, the South African technology services firm identifies three recurring gaps that emerge as organisations scale their platforms and introduce AI workloads

CASA Software, a South African technology services and solutions company that works with enterprise clients across finance, telecommunications, retail and the public sector, has published an analysis identifying the operational blind spots most commonly encountered by large organisations during technology transformation programmes.

The analysis draws on findings from Broadcom and Gartner to frame three recurring challenges. The first is the treatment of artificial intelligence as a standalone initiative rather than a shared platform capability. Broadcom has projected that during 2026, AI will move beyond isolated pilots to become embedded across enterprise platforms, supporting application development, infrastructure operations and data processing workflows. Gartner adds that by 2028, approximately 70% of enterprises will have integrated AI-augmented software testing tools into their engineering toolchains, up from around 20% in early 2025.

"Transformation is no longer about doing more. It is about removing the blind spots — technical, operational, and economic — that slow organisations down," said Michael Brink, Chief Technology Officer at CASA Software. "Enabling faster innovation without increasing the operational burden for IT has become a leadership challenge. This next phase of transformation isn't about adding more technology; it's about identifying and addressing the blind spots that slow progress as environments scale, platforms multiply, and operational complexity grows."

The second gap relates to data governance and regulatory compliance in AI deployments. As governments introduce stricter data residency requirements and AI accountability frameworks, organisations that have not accounted for these constraints at the platform design stage face the risk of having to rebuild or duplicate infrastructure. Brink described this as a growing assumption problem: that AI can be operationalised independently of where data resides and how it is regulated. "Enterprises need platforms that support Sovereign AI by design. Such models allow data pipelines to run locally without fragmenting development or operations, enabling innovation while meeting sovereignty, regulatory, and cost constraints," he said.

The third issue involves infrastructure fragmentation. Many large organisations built their technology environments by selecting separate best-of-breed tools across compute, storage, networking and security. While each component may perform well individually, operating them as disconnected stacks creates friction that slows delivery as environments scale.

Broadcom's position, which Brink endorsed, is that the challenge is not best-of-breed technology itself but the absence of a unifying operating model across those components. "Organisations will increasingly standardise on unified platforms that integrate these capabilities. Hyper-converged, software-defined infrastructure simplifies the foundation, but agility is delivered by what sits above it: shared lifecycle management, integrated Kubernetes services, consistent security policies, and automated operations," said Brink.

On security, Brink argued that protection needs to be embedded at the platform layer rather than applied as a separate process. "When security is embedded in the platform, teams spend less time coordinating controls and more time delivering outcomes. Protection becomes part of the platform's default behaviour rather than a separate process, allowing organisations to move faster while maintaining strong security and compliance," he said.

Broadcom's assessment, cited in the analysis, is that enterprise transformation programmes rarely fail because of a lack of ambition or technology, but rather when accumulated complexity begins to undermine agility and operational efficiency. "Organisations that make progress recognise these blind spots early and address them through deliberate platform design," said Brink.