Artificial intelligence has rapidly moved from boardroom curiosity to boardroom mandate. Across South Africa — from Sandton’s financial district to Cape Town’s technology hubs — executives are under growing pressure to demonstrate how AI will deliver real business value.
Yet amid the surge in investment and experimentation, a critical question remains: where is the return on investment?
For many organisations, AI pilots are multiplying but measurable outcomes remain elusive. The challenge facing today’s C-suite is no longer whether to adopt AI, but how to translate its promise into tangible results.
The missing ROI
There is a growing disconnect between boardroom expectations and operational outcomes. While organisations are experimenting widely, only a small percentage have succeeded in translating AI initiatives into meaningful financial impact.
For South African CFOs and CIOs operating in a constrained, low-growth economic environment, this is a critical warning. Capital cannot be absorbed by experimental projects that never integrate into core business processes.
So, why are so many well-resourced organisations struggling to unlock value?
Part of the answer lies in how the technology is contextualised. Despite its remarkable generative capabilities, AI is fundamentally another technology — albeit a powerful, general-purpose one. Like cloud computing, mobile platforms and enterprise software before it, AI adoption is subject to the same organisational, economic and behavioural forces that shape every major technology shift.
The latest Gartner “hype cycle” helps explain the current moment. Generative AI — the dominant technology story of the past two years — has passed what Gartner calls the “peak of inflated expectations” and is entering the “trough of disillusionment”.
A practical philosophy: humans and AI together
Despite the pessimistic name, this phase is not a failure. It is a necessary stage of technological maturity, during which initial excitement gives way to the practical realities of implementation: cost management, data governance, security, regulatory compliance and — most importantly — demonstrable business value.
For South African executives, the shift is good news. The focus is moving away from novelty toward identifying targeted, high-impact use cases that deliver measurable outcomes.
Successfully navigating this transition requires a clear philosophy: the future of work depends on humans and AI working together, not AI replacing people.
This distinction is particularly important in the South African context, where skills development, employment and human capital growth remain critical national priorities.
Enterprise platforms such as Workday have spent more than two decades helping organisations manage major technological shifts while maintaining strict standards around data security, reliability and compliance.
AI’s computational power can dramatically accelerate data analysis, pattern recognition and routine processes. Humans contribute contextual judgment, empathy, creativity and strategic thinking. Combined, these capabilities create significantly more value than either could deliver alone.
Organisations already seeing meaningful AI ROI tend to follow this principle closely. They are not necessarily using better AI models — they are applying them more intelligently: streamlining existing workflows, automating repetitive administrative tasks and reducing operational bottlenecks.
Enterprise platforms such as Workday have spent more than two decades helping organisations manage major technological shifts while maintaining strict standards around data security, reliability and compliance. One insight has emerged clearly during the current AI transition: the most effective deployments use AI to augment human capability, not replace it.
Generative AI tools are helping surface insights from complex datasets, generate first-draft content and accelerate research. Natural language assistants embedded within enterprise systems help employees find information faster and make better decisions.
In practice, these tools function less like replacements and more like cognitive accelerators — giving employees a head-start on complex work.
The rise of the digital worker
As AI capabilities mature, organisations are beginning to integrate a new category of workforce participant: AI agents.
Unlike traditional chatbots or simple automation scripts, AI agents can autonomously perform multi-step tasks to achieve defined goals — reconciling data across systems, executing routine workflows and managing high-volume operational processes.
The result is the emergence of a hybrid workforce composed of full-time employees, contractors and contingent workers, and digital workers powered by AI. By delegating repetitive, lower-value tasks to digital workers, organisations free human employees to focus on higher-value activities such as strategic analysis, complex problem-solving and customer relationships.
But this also introduces the challenge of “agent sprawl” — the inevitable proliferation of agents across the enterprise. Organisations will need centralised governance to onboard AI agents, define their roles and data permissions, and monitor their activities, whether those agents are built in-house or sourced from third-party vendors.
The core value of such governance lies in bringing enterprise-grade accountability and visibility to AI investments, enabling leaders to track agent ROI, enforce data security and compliance standards, and measure operational impact.
It is for this reason that Workday introduced an Agent System of Record (ASOR), a centralised platform designed to help organisations govern, integrate and optimise their growing fleet of AI agents alongside their human workforce. Functioning much like a traditional HR system but tailored for a digital workforce, ASOR provides a single hub to securely onboard AI agents, define their roles and data permissions, and monitor their activities in real time — whether those agents are built by Workday, custom-developed or sourced from third-party vendors.
For organisations, the core value of ASOR lies in bringing enterprise-grade accountability and visibility to AI investments: it enables leaders to track agent ROI, enforce data security and compliance standards, and measure operational impact. By treating AI agents as an integrated part of the overall organisational structure, ASOR empowers businesses to scale agentic AI, drive process efficiencies and manage a blended workforce where digital and human labour collaborate effectively.
Combining deterministic and probabilistic systems
Another concept business leaders must understand is the distinction between deterministic and probabilistic systems.
Deterministic systems follow predefined rules and execute precise instructions. They are essential in environments requiring accuracy and reliability, such as financial transactions, compliance processes and engineering calculations.
Probabilistic AI systems, including large language models, operate differently. They infer patterns from data and generate outputs based on statistical likelihood, allowing them to handle complex language, ambiguous information and large unstructured datasets.
Each has strengths and weaknesses. Deterministic systems can be rigid and struggle with messy, unstructured inputs. Probabilistic systems, while flexible, can produce incorrect outputs — so-called hallucinations.
The most effective enterprise solutions combine both approaches, often referred to as hybrid AI or neuro-symbolic AI, integrating rule-based precision with probabilistic reasoning to create platforms that are both reliable and adaptable.
A C-suite playbook for AI ROI
For South African organisations determined to move beyond experimentation and achieve measurable ROI, AI adoption must be led across the entire executive team.
- CEO: set the strategic direction. The CEO must shift the organisational narrative from “we need AI” to “we need to solve this specific business problem”. Successful leaders position AI as a tool for solving defined operational or strategic challenges while reinforcing that AI is designed to augment employees, not replace them.
- CFO: enforce ROI discipline. AI investments should be treated like any other capital expenditure. CFOs must require clear success metrics before projects begin. Whether the goal is reducing financial close cycles, improving customer acquisition efficiency or lowering operational error rates, outcomes must be measurable and continuously monitored.
- CIO: secure and strengthen data foundations. AI is only as effective as the data it uses. CIOs must ensure enterprise data is clean, structured, accessible and secure. In South Africa, this also means ensuring compliance with the Protection of Personal Information Act while enabling responsible innovation.
- CHRO: prepare the hybrid workforce. AI adoption is fundamentally a workforce transformation challenge. CHROs must invest in upskilling, change management and workforce redesign. Employees need training on how to collaborate effectively with AI tools and manage digital workers.
- COO: embed AI into operations. Rather than pursuing massive system overhauls, COOs should focus on embedding AI into existing workflows where it can remove friction and deliver immediate efficiency gains. Incremental integration often produces faster and more sustainable ROI than large-scale transformation initiatives.
Moving beyond the hype
The era of inflated expectations around AI is fading — and for South African organisations, this is a moment of opportunity. The race to adopt AI simply to claim participation is ending. In its place is a more mature phase focused on practical implementation, disciplined investment and measurable outcomes.
The organisations that succeed will not necessarily be those with the most advanced algorithms, but those that integrate AI most thoughtfully into their operations — and finally unlock sustainable, measurable ROI.