Building AI Capability Across the Workforce

What It Really Takes to Make AI Work

AI is not waiting for organisations to be ready.

It is already embedded in how work is getting done. What is less visible is how unevenly it is being used. In many organisations, capability is developing quietly. One team experimenting confidently. Another avoiding it entirely. Often within the same function.

This is the reality organisations are navigating.

Organisations are not struggling to adopt AI. They are struggling to design for human capability in an environment that is moving faster than their learning models.

Data from the EY Australian AI Workforce Blueprint, captured in 2025, provides a useful snapshot of this gap. At that point, 68 percent of computer based workers were already using AI, yet only 35 percent had received formal training. At the same time, 72 percent expressed concern about breaching data or regulatory rules, and 42 percent had not been given a clear reason or purpose to use AI in their role.

Further, up to 40 percent of potential productivity gains were being left unrealised due to gaps in workforce capability strategy .

That was the picture then.

It is unlikely to have softened. Adoption will have accelerated. Expectations will have increased. The gap between confident and hesitant users will have widened.

The gap between access and capability is widening, not narrowing. And unless organisations intentionally design for it, that gap will become one of the defining performance divides of the next decade.

This is not a technology problem.

It is a design problem.

Confidence is the missing layer in most AI capability strategies. Organisations are investing in tools and training, yet overlooking the psychological shift required for people to trust, apply, and experiment with AI in real work. Without confidence, capability stalls regardless of investment.

Case Studies: What Leading Organisations Are Doing Differently

Commonwealth Bank: Designing Capability as a System

Commonwealth Bank’s investment in AI capability exceeds 90 million dollars, with more than 30,000 employees completing AI learning by early 2026. While the scale is significant, the real insight lies in the structure.

Learning is tiered. Foundational programs establish awareness. Practical workshops focus on applying AI to real operational tasks. Advanced programs build deeper capability for leaders and technical teams. AI tools are embedded directly into workflows rather than positioned as separate learning experiences.

The CEO described this as a structural redesign of how skills are developed and how careers evolve, rather than a training initiative (Source: AFR / colitco.com, 2025–2026) .

This matters because it reframes learning from instruction to experience and that shift is where capability accelerates.

Coles and Wesfarmers: Relevance at Scale

The OpenAI for Australia initiative, involving organisations such as Coles and Wesfarmers, is targeting AI literacy at scale across more than 200,000 employees.

What stands out is not just the reach, but the relevance. Learning is designed around real roles in retail and operations, not generic AI concepts.

Capability does not build through abstraction. It builds when people can see direct application in their own work.

These organisations demonstrate that frontline capability can be developed at scale when it is grounded in context and relevance (Source: openai.com/global-affairs/openai-for-australia, December 2025) .

PwC: Building Confidence Through Experience

PwC’s My AI program has reached over 75,000 employees globally, with 95 percent engaging with generative AI.

A defining element of this program is the use of facilitated team sessions where employees experiment with AI on real work challenges. These sessions remove the pressure of formal learning and replace it with structured exploration.

This matters because it reframes learning from instruction to experience and that shift is where capability accelerates.

Capability builds faster when people are actively using AI, reflecting on outcomes, and learning together. It becomes social, visible, and embedded (Source: hrexecutive.com / pwc.com) .

Suggested Methods: Approaches That Actually Work

Across these examples, capability is not being built through a single intervention. It is being shaped through a set of deliberate methods that reflect how people learn and work.

Prompting Sessions Short, facilitated sessions where teams work with AI on real tasks. These sessions build shared language, reduce hesitation, and create immediate relevance.

AI Use Case Sprints Focused workshops to identify where AI can improve quality or efficiency in specific workflows. This ensures capability is aligned to real business needs.

Capability Tiering Defining levels of capability across the workforce. For example: Awareness for all. Application for those using AI regularly. Advocacy for those supporting others.

Learning in the Flow of Work Embedding prompts, examples, and guidance into existing tools and systems. Capability develops when learning is integrated into daily activity, not separated from it.

Peer Champion Networks Identifying early adopters and enabling them to support others. Capability spreads faster through peer influence than through formal instruction alone.

Responsible Use Frameworks Providing clear guidance on appropriate use. Confidence increases when people understand the boundaries within which they can operate.

These methods are not complex.

Their effectiveness lies in their alignment to how capability actually develops. Through use, interaction, and reinforcement.

A Practical Starting Point: Regardless of Your Organisation’s Size

For organisations beginning or refining their approach, progress does not require large scale investment. It requires intentional design.

Start with a real problem. Identify a workflow where AI can deliver immediate value. Relevance builds momentum.

Define capability levels. Clarify what awareness, application, and advocacy look like in your context.

Run an initial experimentation session. Observe how people engage, where they hesitate, and where they find value.

Establish governance early. Address uncertainty directly to build confidence.

Leverage early adopters. Identify those already using AI effectively and involve them in supporting others.

Clarify ownership. Determine how HR, L&D, IT, and the business will collectively support capability development.

The Shift That Matters

The organisations that will realise the value of AI are not those that move fastest.

They are the ones that design most intentionally for human capability.

The question for HR and L&D is not whether to respond. It is how deliberately they are choosing to design for this shift. Because in the end, the differentiator will not be artificial intelligence. It will be human capability.

Reference Sources (They all open in a new tab)

EY Australian AI Workforce Blueprint 2025 https://www.ey.com/content/dam/ey-unified-site/ey-com/en-au/services/ai/documents/ey-au-ai-workforce-blueprint-2025-final.pdf

EY Work Reimagined Survey 2025 (global, 15,000 employees) https://www.ey.com/en_gl/workforce/work-reimagined-survey

Commonwealth Bank Future Workforce Initiative (AFR, 2025–2026) https://colitco.com/commonwealth-bank-launches-a90m-ai-upskilling-program/

OpenAI for Australia, Coles, Wesfarmers, CommBank (Dec 2025) https://openai.com/global-affairs/openai-for-australia

PwC: Unleashing Gen AI, My AI Program (2023–2025) https://hrexecutive.com/heres-what-a-1-billion-investment-in-gen-ai-can-really-deliver/


Vanessa Trower is the Founder and Principal Consultant of Nexperk, a Sydney-based people and performance partner for large frontline workforces. She hosts The Transformation Edit and SHIFT Happens on the Front Line podcasts. Want to talk about what this could look like in your organisation? Book a chat with Nexperk.

Previous
Previous

Reframing Frontline Onboarding in Retail: A Confidence-Led Model for Real-World Readiness

Next
Next

“What Prompt Did You Use?”