While much of the AI-and-jobs debate centres on which roles artificial intelligence will replace, the reality is different: AI is not coming for your team; it is coming for outdated hiring strategies that focus on simply filling a role.
Demand for AI-related skills in South African job postings grew 77% year-on-year in 2025 and by 352% since 2019, according to Pnet’s Job Market Trends Report 2025.
The supply of talent cannot even come close to keeping pace.
Some may simply call this a shortage, but I would call it a signal: the gap points to a deeper problem in how skills are defined, sourced, and deployed.
A gap that wide, for so long, is no longer a pipeline problem; it is a model problem — an indication that the fundamental way in which skills are defined, sourced and deployed needs to change.
Context clues
To understand where it falls short, consider how the AI hiring model typically plays out in a financial services context.
A bank decides to embed AI into its credit decisioning process and appoints a data scientist, a machine learning engineer, and an AI implementation lead.
Months later, the anticipated value has still not materialised, because the necessary capability is lacking.
The core issue is often the absence of a clear answer to a fundamental question: “What problem are we actually solving, and what does success look like?”
In my experience partnering with technology teams in financial services, the main challenge is not always a shortage of AI talent; rather, it is a lack of structure, clarity, and alignment.
Initiatives stall because data governance standards are unclear, automation runs on an outdated operating model, and teams are resistant to new AI tools due to insufficient change management or capability transfer.
Even when talented people are in the room, these issues manifest as burnout, rework, missed regulatory deadlines, and slow delivery, even if they do not typically appear in skills audits.
In reality, AI doesn’t just change which skills are required; it changes the nature of work itself.
Companies are no longer hiring for a role. They’re trying to build a system.
They’re trying to build a system.
Guidance
In my experience, an effective way to approach this new objective is to follow three steps.
The first is defining the expected outcome before specifying the role.
This is a leadership discussion, focused on what is needed for success.
For example, if a bank wants to embed AI in its credit decisioning, this step might involve agreeing in advance on the operational definition of embedded AI, the data quality standards, the model governance thresholds, and what success looks like.
Without this clarity, even highly qualified hires lack direction.
Next, reframe staff augmentation as a long-term solution.
Structured correctly, it enables access to scarce skills, facilitates knowledge transfer to internal teams, and sustains delivery momentum while building long-term capability.
The most successful organisations design staff augmentation engagements with clear outcomes, accountability and detailed capability transfer to ensure lasting value.
A third step is to understand and accept that AI readiness is not a hiring event.
It is a continuous organisational capability that requires structured governance, clear data foundations, and leadership alignment that no single appointment can provide.
The financial institutions making the most measurable progress on AI are those that have created the conditions for AI to function, not necessarily those with the largest AI headcount.
Opening another AI role? Three questions to ask first:
- What specific outcome is this hire accountable for, and how will we measure it?
- Do we have the surrounding structure for this role to succeed: clean data, governance, and a change process?
- Could external staff augmentation deliver the same outcome faster, while also building internal capability?
Case studies
External augmentation partnership models have already proven effective in South Africa.
For example, financial services provider Discovery Bank built its AI-powered infrastructure through external technology partnerships, while GoTyme Bank developed its entire banking platform through an external partnership model.
These examples show that South African financial institutions can move faster and more decisively when they treat external expertise augmentation as a strategic capability rather than a temporary solution.
AI will widen the gap between institutions that have built the right foundations — clear governance, integrated capability, and outcome-linked roles — and those that haven’t.
In financial services, the window to build those foundations correctly is narrower than in other industries, because regulatory requirements, legacy core systems, and complex data environments constrain the pace of AI adoption.
The organisations that will succeed in the age of AI will not be those with the largest AI teams. Rather, they will be those with a workforce strategy focused on continuous transformation and the right AI foundations for talent to deliver clear outcomes.
This is what it means to build a system, and it is the hiring strategy suited for the AI-driven future.
