The intelligent workplace (part 2): Technology’s next transformation of work

Part 2: Managing performance in a human-AI workforce

Managers have traditionally organized work around people: assigning responsibilities and holding individuals accountable for results. AI changes that model. When employees delegate tasks to AI assistants and teams include autonomous agents, performance emerges from a system of people, technology, data, processes, and managerial choices.

Part 1 of this series explored the rise of the intelligent employee experience and how AI assistants and connected workplace platforms are changing everyday work. However, as intelligent systems take on a more active role, organizations must reconsider not only how work is performed, but how it is managed and measured.

Part 2 examines what happens when AI becomes a permanent member of the team. It considers how leaders must manage blended human-AI workflows and how employee development can preserve judgment and accountability as routine tasks move to intelligent systems.

Microsoft reports that 36% of managers expect to supervise AI agents within five years. Leaders also anticipate that teams will build multi-agent systems and redesign processes. Nearly one-third expect to hire AI agent specialists within 12 to 18 months, while 28% are considering specialist AI workforce managers.

Yet adding digital workers does not automatically create business value. Only 14% of workers were using generative AI daily in 2025, according to PwC, while 56% of CEOs said their AI investments had delivered neither revenue growth nor cost savings. The gap suggests that managing a human-AI workforce is not primarily a software challenge. It requires organizations to rethink how people develop expertise, and where accountability rests when machines influence decisions.

Performance moves beyond output

Traditional performance systems often reward visible activity: cases closed, documents produced, calls handled, or hours billed. AI can increase all these numbers without necessarily improving the work. A poorly designed system can generate reports, code, marketing copy, or customer responses at extraordinary speed, but greater volume has little value if people must correct inaccuracies or repair damaged trust.

Dr. Janet Bastiman, chief data scientist at Napier AI, says quality should be the target rather than volume. “It would be trivial to automate slop, and that never results in a good company outcome,” she argues. Performance measures should instead connect work to organizational purpose and assess effectiveness and accuracy, regardless of whether AI was involved.

Good performance increasingly includes framing the right problem and accepting responsibility for the result. The most productive employee may be the one who prevents a plausible AI error from reaching a client or helps develop a more reliable workflow.

Margarita Lindahl, head of AI at Panasonic Connect Europe, explained: “The strongest performers will not necessarily be those who produce the most with AI. It will be those who use it to create better decisions, stronger customer outcomes, and build knowledge, ultimately benefiting the wider organization.”

Existing performance practices may not be ready for this change. Only 48% of UK employees report even a basic level of formal performance management, defined as having specific objectives for their roles, according to the CIPD. Meanwhile, nearly half of employees say their work feels chaotic and fragmented. Measuring tasks and activity in that environment risks confusing busyness with contribution.

Performance frameworks must combine output with quality, judgment, collaboration, customer value, and responsible AI use. Results also depend on suitable tools, reliable data, training, and time for human review—not talent alone.

These conditions connect directly to the intelligent employee experience covered in Part 1 of this series. Performance cannot be separated from the environment in which employees work. Fragmented platforms and technology that add friction will affect results, regardless of how sophisticated an organization’s performance measures become.

Managers become architects of human-AI teams

AI expands rather than removes managerial responsibility. Systems may allocate tasks, recommend priorities, monitor work, or provide performance insights, but leaders still decide how they operate and when employees can override them. Management becomes less about controlling activity and more about designing complementary human and machine capabilities.

“Business leaders need to increasingly become architects of human-AI systems,” Lindahl tells ITPro. That requires enough technological literacy to distinguish genuine capability from hype, where human judgment remains essential, and to establish clear accountability. As AI influences more decisions, she says, leaders must create responsible guardrails rather than surrender authority to the technology.

Deloitte found that managers spend almost 40% of their time resolving immediate problems and completing administration, but only 13% developing people. More than one-third feel unprepared for the people-management aspects of their roles. AI could return time to coaching, or create another stream of alerts and scores.

Andrew Avanessian, CEO of Haiilo, emphasized that leaders will need stronger systems thinking because AI often exposes inefficiencies elsewhere rather than solving them. They must also communicate what is changing, why it is changing, and where employees continue to create value.

Managers must also preserve permission to challenge the machine. If an agent allocates a task or recommends an action, employees should know whether they are expected to follow or overrule it. Otherwise, accountability becomes blurred, and people may defer to a system simply because its recommendation appears authoritative.

Development must protect human expertise

As AI absorbs routine and analytical tasks. The larger challenge is ensuring that people still acquire the domain knowledge and practical experience needed to evaluate machine output.

Employers expect 39% of workers’ core skills to change by 2030, according to the World Economic Forum. If the global workforce were represented by 100 people, 59 would require training, yet 11 may not receive it. Although 85% of employers plan to prioritize upskilling, only 33% of UK businesses using or considering AI say they are training or retraining existing employees in AI-related skills.

This gap particularly threatens early careers. Routine work has served as an apprenticeship: junior employees gather information, produce drafts, check details, and observe how experienced colleagues turn evidence into decisions. AI may remove those activities, but not the need for underlying knowledge.

Jenny Briant, director of talent Strategy at Scale Factory, explained that employers must create deliberate opportunities for employees to practice, make decisions, and receive feedback. “Employers cannot assume that people will develop judgement simply because they have access to better tools,” says Briant. “They need to create deliberate opportunities for employees to practise, make decisions and receive feedback. Practitioner-led training, mentoring and supervised work on real problems will become more important because they connect technical knowledge to the situations employees actually face.”

The goal is not to make people compete with machines on speed. It is to strengthen critical thinking, problem framing, creativity, relationship-building, and contextual judgment. Bastiman emphasized, “Ensuring meaningful human oversight of AI systems through various types of checks will ensure that the employees keep their core skills sharp for the task, but allowing the shift of work into new tasks will allow the building of new skills that were not possible prior to the overload of routine tasks.”

Employee development must also include the ability to explain how AI contributed to an outcome. Asking someone why they selected a tool, how they checked its response, and which elements required personal judgment reveals more about professional competence than banning the technology during an assessment.

Developing this combination of technological fluency and human judgment will become increasingly important as organizations prepare for 2030. Part 3 of this series examines how skills-based workforce planning and internal mobility can help businesses build the capabilities required as AI, automation and other emerging technologies reshape work.

Trust determines whether measurement improves work

AI promises more personalized and evidence-based performance management. It could identify patterns that a busy manager might miss or reveal where workflows routinely break down. Yet the same technology can intensify surveillance and give questionable judgments a veneer of objectivity.

Algorithmic management is already widespread. OECD research found that such tools were used to instruct, monitor, or evaluate workers in 90% of surveyed US firms and an average of 79% across France, Germany, Italy, and Spain. Managers identified unclear accountability, limited explainability, and insufficient protection of employee health among their concerns.

Bastiman warns that automated evaluation can remove empathy from performance management and embed assumptions from the culture in which the system was developed. Briant similarly notes that an algorithm may count output or response times but miss an employee who supported a struggling colleague or challenged a poor decision. What is easiest to measure is not always what is most valuable.

Transparency and consultation are therefore operational requirements, not optional ethics statements. Employees should understand what data is collected and how they can challenge it. OECD evidence indicates that training and consultation are associated with better outcomes when workplace AI is introduced. Trust matters because only 58% of workers currently trust their direct manager and feel able to speak openly with them, according to PwC.

Organizations that manage human-AI performance successfully will redesign work, protect time for learning, measure the value of outcomes, and keep accountability visibly human. AI can expand capacity, but leadership determines whether it produces better decisions or simply more activity. Performance will depend on how well people and machines work together, and whether employees emerge more capable and responsible.

Managing today’s human-AI workforce also requires organizations to prepare for the more extensive transformation ahead. The final part of this three-part series looks toward the workplace of 2030, identifying the emerging technologies and workforce strategies that will determine whether businesses can turn continual disruption into lasting competitive advantage.