There is a significant idea emerging from the rapid development of artificial intelligence: the value of simply knowing how to operate software is changing, while the value of knowing what should be done, why it matters and whether the result is fit for purpose is increasing.

Headlines such as "The Keyboard Is Done" are deliberately provocative. Keyboards are not disappearing overnight, and technical skills will not suddenly become worthless. Beneath the sensational language, however, lies a more important development: the way knowledge work is organised, performed and managed is beginning to change.

For decades, professionals have created value partly through their ability to operate digital tools. Artificial intelligence is changing this relationship and challenging organisations to reconsider how work is designed, how responsibilities are assigned, how decisions are made and how performance is measured.

1. From Operating Software to Directing Intelligence

A financial professional might spend hours building spreadsheets, applying formulas, preparing reports and creating charts. A strategy consultant might spend days gathering information, reviewing documents, conducting analysis and preparing recommendations. A manager might spend considerable time compiling reports, following up on actions and preparing performance updates.

Traditionally, software helps these professionals perform the work, but they remain responsible for operating the tools and coordinating the steps. AI can increasingly assist with multiple parts of these activities. More advanced AI agents can potentially plan and perform sequences of tasks, interact with supported software and information sources, analyse data and execute authorised actions within defined boundaries.

This does not eliminate the need for professional expertise. It changes where that expertise creates the greatest value. The critical question becomes less about whether someone knows how to operate a particular application and more about whether they understand the business problem well enough to direct the technology, evaluate its output and determine what should happen next.

2. The Value of Technical Skills Is Changing, Not Disappearing

Technical skills will remain important. Financial modelling, data analysis, coding, presentation design and software proficiency will continue to matter in many professions. However, simply performing predictable digital tasks may become less differentiating as AI capabilities improve.

The more important shift is towards combining technical competence with broader professional capabilities.

The implication is not that execution becomes unimportant. Rather, execution alone may become less differentiating. Professionals who combine domain knowledge, judgement, AI capability, leadership and decision-making can potentially produce better work and contribute more directly to organisational outcomes.

For organisations, this raises an important question: are employees being trained merely to use AI tools, or are they developing the capability to redesign work and improve results using AI?

3. The Rise of AI Agents Changes the Unit of Work

Generative AI has made it possible to request summaries, analyses, drafts and recommendations through conversational interfaces. AI agents extend this concept towards systems that can potentially plan and carry out sequences of actions, use tools and interact with business systems within specified permissions.

An AI assistant might explain how to prepare a performance report. A more capable agent might gather authorised data, analyse performance against agreed measures, prepare a draft report and identify issues requiring management attention.

Instead of treating every task as something a person must perform manually, organisations can increasingly examine the outcome required and determine which combination of people, AI assistants, agents and automation is appropriate.

The answer depends on complexity, risk, judgement requirements, data quality, customer impact, regulatory obligations and the consequences of an error. Not every process should be automated, and not every decision should be delegated to AI.

4. Human Judgement Becomes More Important, Not Less

AI can produce an impressive business plan, strategy, financial model, policy, dashboard, proposal or presentation. The quality of the presentation, however, does not establish the accuracy of its contents. AI can make mistakes, misinterpret context, rely on incorrect assumptions or produce plausible but unsupported conclusions.

Four questions should guide AI-enabled work:

• Outcome: What exactly should be produced or achieved?
• Context: What information, assumptions, policies, standards and constraints must the system understand?
• Control: What can AI do independently, what requires approval, and what must remain human-led?
• Quality: How will the organisation establish whether the result is accurate, reliable and fit for purpose?

These questions should become part of the organisation's operating discipline, not merely instructions given to an AI tool. AI can contribute analysis and execution, but people remain accountable for the decisions assigned to them.

5. Beyond Prompt Engineering: Designing AI-Enabled Work

Prompt engineering is useful, but knowing how to write prompts is only one part of using AI effectively. A deeper capability is AI work design: determining how tasks, decisions, controls and responsibilities should be distributed between people and intelligent systems.

1. Define the objective and business problem.
2. Specify the required outcome and quality standards.
3. Provide relevant context and reliable information.
4. Establish constraints, permissions and decision rights.
5. Determine which activities are suitable for AI assistance or agent execution.
6. Allow the system to perform the authorised work.
7. Verify the output against agreed standards.
8. Obtain human approval for consequential decisions.
9. Measure results and capture lessons.
10. Improve the workflow as evidence and capabilities evolve.

This approach moves organisations beyond isolated productivity experiments towards repeatable, controlled and measurable AI-enabled processes. Adding an AI tool to an inefficient process may accelerate the wrong activities. Redesigning the process can create a more meaningful improvement.

6. What This Means for Organisational Performance

The implications extend across the organisation. Jobs, workflows, structures, performance systems, management practices, capability requirements and governance arrangements may all need to evolve.

Introducing these technologies does not automatically improve organisational performance. Organisations must still establish clear objectives, assign responsibilities, align resources, define performance measures and manage risks. They must also determine whether the change actually improves productivity, quality, cost, cycle time, customer outcomes or other relevant measures.

7. The Opportunity to Redesign How Organisations Work

Organisations that approach AI strategically have an opportunity to rethink how work is organised rather than simply adding new tools to existing processes. This requires examining the relationship between business objectives, operating processes, people, technology, decision rights and performance expectations.

For example, a business may introduce AI to accelerate customer service reporting. A narrow approach would focus on generating reports faster. A broader approach would examine whether the system can also identify recurring customer complaints, reveal underlying process problems, direct issues to the right teams and help managers assess whether corrective actions are effective.

The second approach focuses on the business outcome rather than the technology alone. The same thinking applies to finance, human resources, procurement, logistics, marketing, project management and executive decision-making.

The objective should not be to automate every activity. It should be to determine where intelligent systems can improve results, where human judgement is essential and how the two can work together effectively.

8. The Future of Digital Business Platforms

AI agents create opportunities for digital platforms to evolve beyond traditional data entry, dashboards and reporting. With appropriate integrations, reliable data and suitable governance, future platforms could help leaders answer four practical questions:

• What is happening?
• Why is it happening?
• What is likely to happen next?
• What action should we consider?

For example, an executive performance assistant could identify a delayed strategic initiative, explain contributing constraints based on available evidence, flag overdue actions and recommend management interventions for review.

Such a system would go beyond reporting. It would support decision intelligence by connecting performance evidence with the decisions leaders need to make.

These capabilities require careful design. They depend on reliable data, clear business rules, appropriate system permissions, human oversight and the ability to explain and verify recommendations.

9. The Future of Professional and Advisory Services

AI can support research, analysis, document processing, modelling, drafting and routine administration. This can create opportunities for professionals to spend more time understanding client problems, interpreting evidence, developing recommendations and supporting implementation.

However, AI-generated outputs are not a substitute for professional accountability. Clients still need advice grounded in their circumstances, sound judgement, appropriate verification and a clear understanding of risks and consequences.

The strongest future model is one in which professional expertise and intelligent systems complement each other. For advisory firms, this creates an opportunity to combine human expertise, digital tools and repeatable processes to improve the quality, speed and consistency of client delivery.

10. Who—or What—Should Do the Work?

Organisations will need a deliberate way to determine how different activities should be performed. For each significant process, leaders should consider whether the work should be human-led, human-led with AI assistance, supported by an AI assistant, performed by an AI agent within defined permissions, or automated through conventional systems.

The decision should reflect complexity, risk, business value, data availability, judgement requirements, regulatory obligations and customer impact.

The objective is not to automate everything. It is to design the most appropriate combination of human expertise, intelligent systems and accountable management for each process.

11. The Productivity Gap Between AI-Enabled and Non-AI-Enabled Work

Employees and teams with effective AI-enabled workflows may be able to produce more—or better—work than those relying entirely on manual processes. A professional who combines domain expertise with AI research, analytical support, reporting automation and governed digital workflows may complete certain activities faster or devote more time to higher-value decisions.

Productivity gains are not automatic. They depend on the task, the quality of the tools and data, the effectiveness of the workflow, employee capability and the quality of verification.

AI capability should therefore be developed alongside process redesign, management capability and performance measurement. Tool access alone is not a reliable measure of transformation.

12. What Leaders Should Do Now

1. Map the work. Identify major processes and knowledge-intensive activities across the organisation.
2. Identify opportunities. Determine which activities could be accelerated, augmented or automated.
3. Redesign workflows. Examine whether the process itself should change.
4. Establish decision rights. Define what people and AI systems may recommend, access, approve or execute.
5. Build capability. Develop judgement, domain expertise, AI literacy and the ability to design AI-enabled work.
6. Establish governance. Protect confidential information, data, intellectual property and accountability.
7. Measure outcomes. Track quality, productivity, cycle time, cost, adoption and business performance.
8. Continuously improve. Use implementation evidence to refine workflows, controls and capability.

These actions should be connected to business objectives and performance priorities. Otherwise, AI adoption risks becoming a collection of disconnected experiments rather than a coherent transformation programme.

Conclusion: From Software Users to Intelligent Organisations

The claim that "the keyboard is done" should not be interpreted literally. Keyboards will remain. Software will remain. Technical skills will remain valuable. What is changing is the relationship between people, technology and work.

Value is increasingly moving beyond simply knowing how to operate technology towards understanding what should be done, why it matters, how work should be structured, what AI can appropriately perform and whether the result is good enough.

For individuals, this means developing judgement, domain expertise, problem formulation, AI literacy and leadership. For organisations, it means redesigning work, clarifying decision rights, building capability and connecting AI-enabled processes to measurable business outcomes.

The future organisation will not necessarily be the one that uses the most AI tools. It will be the one that combines human judgement, organisational knowledge, artificial intelligence, digital systems, disciplined execution and performance management most effectively.

That is a much deeper transformation than simply adopting a new AI application. It is a transformation of how work is designed, how decisions are made and how organisations perform.

The real competitive advantage will come from connecting human judgement, organisational knowledge, artificial intelligence, digital systems, disciplined execution and performance management to produce measurable results.