For much of the discussion around artificial intelligence and the workforce, the dominant question has been whether AI will replace jobs. It is an understandable concern, but it may not be the most interesting workforce question facing organisations over the next few years. A potentially more significant change is already emerging: what happens when AI enables an average employee to perform much more like a very good one?
The distinction matters because the impact of AI may not initially be characterised by widespread job replacement. Instead, it may appear through a gradual compression of capability across the workforce. Tasks that once required significant experience, specialist knowledge or simply more time can increasingly be supported by technology that can analyse information, structure an argument, create a presentation, draft correspondence, summarise complex material and suggest solutions in seconds.
This does not mean that AI suddenly gives someone ten years of experience. It does, however, mean that some of the visible differences between employees may become considerably smaller.
Research is beginning to point in this direction. One widely cited study involving more than 5,000 customer service agents found that access to generative AI increased productivity overall, but the largest improvements occurred among less experienced and lower-performing workers. Similar research involving consultants found that AI improved both the speed and quality of work, with below-average performers experiencing particularly significant gains.
The implication is worth considering. AI may raise the floor of workforce performance before it meaningfully raises the ceiling.
When competence becomes easier to achieve
Organisations have traditionally differentiated employees through a combination of knowledge, experience and execution. One employee may write better reports, another may analyse information more effectively, while someone else may know how to structure a difficult problem or communicate an idea clearly to senior stakeholders. Over time, those differences contribute to how organisations identify high performers, allocate opportunity and make decisions about progression. AI complicates that equation.
An employee who previously struggled to create a well-structured document can now generate a strong starting point within minutes. Someone unfamiliar with a particular topic can rapidly develop a reasonable understanding of it. A junior employee can use AI to structure an analysis in a way that previously may have required support from someone significantly more experienced. The underlying expertise has not necessarily changed, but the quality of the observable output may have.
That creates what could be described as capability compression: the gap between acceptable and very good execution begins to narrow because technology gives more people access to methods, information and structures that were previously unevenly distributed across the workforce.
This does not eliminate high performance. It changes where high performance is found. As management thinker Peter Drucker famously observed, “There is nothing so useless as doing efficiently that which should not be done at all.”
That distinction becomes increasingly relevant in an AI-enabled workplace. If technology makes execution faster and competence more accessible, the real organisational advantage shifts away from simply doing work efficiently and towards determining what work is worth doing in the first place.
The performance bar will move
Whenever technology makes an activity materially easier, organisational expectations rarely remain unchanged. The spreadsheet did not result in finance teams performing the same number of calculations and using the remaining time elsewhere. Email did not simply allow organisations to send the same volume of correspondence faster. Technology changes what organisations consider normal. AI is likely to follow the same pattern.
If producing a competent first draft of an analysis once required several hours and can now be achieved in a fraction of that time, the organisation will eventually stop considering the production of that draft to be particularly valuable. The baseline expectation will simply move. Employees may be expected to produce more, respond faster or manage a broader scope. More importantly, however, the value of the employee may increasingly be judged by what happens after the initial output is created.
Did they ask the right question? Did they recognise that the information was incomplete? Did they understand the commercial implications? Did they challenge an assumption that everyone else accepted? Could they distinguish between something that was technically correct and something that was practically useful?
These are fundamentally different measures of performance.
The next performance gap may sit above execution
If AI progressively reduces the gap in basic execution, the characteristics that distinguish exceptional employees may move further up the value chain. Judgement, context, prioritisation, commercial awareness, accountability and influence become more important because they are harder to reproduce simply by generating a better first draft.
Consider two employees who can now both produce an impressive strategic presentation using AI. The presentation itself tells us increasingly little about the capability of the person behind it. The difference emerges in whether the employee understood the problem properly, selected the right information, identified unintended consequences and was prepared to defend the recommendation when challenged.
The same principle applies across functions.
In recruitment, AI can help create job advertisements, identify candidate profiles and draft communications. It cannot fully understand why a technically strong candidate may be entirely wrong for a particular organisation. In supply chain, AI can analyse data, model scenarios and propose plans. It does not automatically understand every operational constraint, supplier relationship or organisational risk sitting behind those numbers. In leadership, AI can draft communications and develop options. It cannot take accountability for the consequences of the decision.
This is where experience may become more rather than less important. Its value simply changes. Instead of experience primarily helping someone produce an answer, it increasingly helps them evaluate whether the answer makes sense.
A polished output is not necessarily capability
There is another organisational risk emerging from this shift. As AI-generated work becomes more sophisticated, it becomes easier to confuse strong output with strong capability.
A polished report can create an impression of expertise. A well-structured recommendation can appear strategically considered. A confident response can sound authoritative even when the underlying reasoning is weak. For managers, this makes the assessment of capability more difficult. The important questions may increasingly sit behind the work rather than inside it. How did the employee define the problem? Which assumptions did they test? What evidence did they consider? What did they reject? Could they explain their reasoning without relying on the generated output? Most importantly, do they understand the consequences of the recommendation they are making?
Organisations that continue to assess performance primarily through visible outputs may therefore find it increasingly difficult to distinguish between genuine capability and AI-assisted competence.
That distinction becomes particularly important when conditions change. AI performs well when the problem is clearly defined and sits within an area where the technology has sufficient context. Real organisational problems are rarely that neat. Information is incomplete, stakeholders disagree, priorities compete and decisions often need to be made without certainty.
That is where human judgement becomes visible again.
The implications for recruitment and development are significant
If AI can accelerate the journey from inexperienced to competent in certain areas, organisations may also need to reconsider what they value when hiring people.
Historically, experience has frequently been used as a proxy for capability. A role might require five years of experience because organisations reasonably assumed that those five years were necessary to develop the knowledge required to perform the work effectively.
That assumption may become less reliable. For some tasks, AI can dramatically reduce the time required to access information, understand a process or create a competent output. The more interesting hiring question therefore becomes less about whether someone already knows how to perform every aspect of a role and more about what they can do with the capability available to them.
Learning agility, curiosity, judgement, communication, commercial thinking and the ability to frame a problem properly may consequently become more valuable than simply having performed the same task many times before.
This does not mean experience becomes irrelevant. It means organisations may need to become considerably more precise about which parts of experience actually create value.
The same applies to learning and development. If AI increasingly provides employees with information and technical support on demand, organisational development may need to focus less heavily on knowledge transfer and more deliberately on judgement, decision-making, critical thinking and commercial understanding.
Managers will need a new definition of high performance
Perhaps the biggest implication sits with leadership. When technology raises the baseline capability of a workforce, managers need to reconsider what they reward. If employees are measured primarily on the volume and polish of what they produce, AI may make performance appear stronger without necessarily making the organisation better. The challenge is to move from measuring outputs to evaluating impact.
A high performer in an AI-enabled workplace may not be the person who produces the most reports, creates the most detailed presentations or responds to the greatest number of requests. It may increasingly be the person who identifies which report does not need to exist, simplifies an unnecessarily complex process, challenges the original question or turns information into a decision that creates measurable value.
That shift has implications for performance management, role design and organisational structure. As routine analytical and administrative work becomes easier to perform, organisations will need to be increasingly deliberate about where accountability sits and what people are actually expected to contribute.
The risk is otherwise straightforward: organisations become dramatically more efficient at producing work without becoming materially better at producing outcomes.
The bigger workforce question
The future-of-work debate is often framed around two extremes. Either AI replaces significant numbers of employees, or it simply becomes another productivity tool that helps people perform their existing jobs more efficiently. The reality is likely to be considerably more nuanced.
Before AI eliminates large numbers of roles, it may fundamentally change what those roles require. The employee does not disappear, but the value they are expected to contribute changes. Entry-level employees may become productive faster. Some experience advantages may narrow. Expectations of output may rise. The value of judgement, context and accountability may increase. Managers may need different ways to recognise genuine capability.
In that environment, the most important question for organisations may not be, “What jobs can AI perform?” It may be: “If AI makes competent performance available to almost everyone, what will differentiate our best people?”
The answer is unlikely to be the ability to create another report, presentation or analysis. It will be the ability to decide what matters, understand why it matters, exercise judgement when the answer is unclear and take responsibility for what happens next. AI may make average employees very good at considerably more things. The organisations that gain the greatest advantage from it will be those that understand that this does not reduce the importance of human capability. It changes where that capability creates value.
The question for organisations, then, is not simply how quickly they adopt AI, but how deliberately they respond to the changes it creates in workforce capability, performance expectations and the value people are expected to contribute.
As AI continues to lift baseline capability across the workforce, the organisations that benefit most will be those that rethink how they define performance, structure roles and develop people. The opportunity is not simply to do more with fewer resources, but to create a workforce that applies better judgement, makes stronger decisions and focuses its effort where it creates the greatest value.
At Australian Project Solutions Group, we work with organisations to strengthen workforce capability, role clarity and organisational design, aligning people, structure and performance with business priorities.
If you are considering how AI is changing the expectations of your workforce, we would be pleased to discuss what that means for your organisation.
Work with the Ones you know.
