Trending
Written by: Izabela Lundberg, M.S.
Why Is A Human Still Carrying Out This Action
This question distinguishes organizations that effectively deploy AI from those that simply purchase it.
Artificial intelligence has created a rather odd issue for business leaders. Although companies are investing heavily in technology that changes how work is done, they tend to use it only to make current work a bit faster rather than asking whether the work should exist in its present form.
This distinction is becoming increasingly important. The talk about enterprise AI has moved from experimentation to implementation, with companies now buying platforms, training staff, setting up governance structures, and putting AI assistants to work across various functions. Yet adopting AI does not guarantee real transformation.
McKinsey's 2025 global study found that 88 percent of those surveyed said they used AI regularly in at least one business function, but only 39 percent observed an enterprise-wide effect on EBIT. Organizations that achieved the greatest value were more inclined to redesign their workflows than to simply add AI to existing processes.
This shows a basic issue with how many organizations approach transformation: they see AI as merely a technological implementation, when the more important question is organizational design.
Suppose a People Operations team is preparing for a new employee's arrival. The onboarding process could involve several people gathering information, sending emails, securing approvals, setting up accounts, updating systems, reconciling records, and then confirming that everything is done. The organization could introduce an AI assistant that drafts emails, summarises policies, and responds to employee questions while still leaving all the original handoffs in place. Efficiency improves slightly, but the overall structure stays the same.
The organization has introduced AI, but it has not restructured its work.
That is why executives should pay much more attention to one question: why is a human still carrying out this task? While the question appears simple, it challenges a core organizational assumption: that current work methods are inherently the way they must continue. This is rarely the case.
Processes accumulate over time. Workarounds become standard practice, spreadsheets become control mechanisms, and manual approvals are added after incidents. Additional reviews are introduced due to past errors, and roles inherit recurring tasks simply because they always have. Years later, people forget the original purpose, but changing the process feels risky.
This is how organizational complexity increases.
AI is revealing this complexity, as many activities considered "work" are actually information transfers, repetitive decisions, administrative coordination, and rule-based tasks. These once required significant human effort, but increasingly do not. The leadership challenge is not to identify where AI fits into the current organization, but to determine how the organization should be structured if AI is available from the outset.
This requires a new conceptual boundary: the Human-Value Line.
The Human-Value Line distinguishes work where human involvement adds unique value from work where it persists only due to inherited processes. Activities requiring judgment, empathy, contextual understanding, ethical reasoning, relationship-building, or accountability fall on one side. Repetitive, predictable, rule-based, administrative, or information-transfer tasks fall on the other.
This boundary is not fixed, nor is it a simple division between "human work" and "machine work." Tasks can shift across the line as technology advances, regulations change, or organizational risk tolerance evolves. The key question is whether leaders intentionally decide where to draw this line.
Too often, they don't make this decision.
The phrase "human touch" often justifies manual processes. However, human involvement is not inherently valuable. Customers do not receive better service simply because employees manually transfer information. Employees do not benefit from onboarding processes that require excessive manual routing. Finance professionals do not exercise better judgment by manually reconciling data that software could handle more accurately.
Human attention is valuable precisely because it is scarce.
When organizations focus attention on tasks that don't require it, they aren't being more human. Instead, they are misusing one of their most valuable resources. This highlights the importance of automation-first design. Traditional process design assumes humans perform the work, with technology as support. In contrast, an AI-native organization assumes processes should run automatically wherever possible, and human involvement should be intentional where judgment, accountability, or relationships improve outcomes.
This approach does not eliminate people from processes. In many cases, it makes their roles more significant. For example, if an automated system manages the administrative aspects of onboarding, collecting information, triggering provisioning, routing approvals, generating documentation, and identifying missing requirements, those who previously coordinated these tasks are not rendered irrelevant. Instead, they can focus on higher-value work.
What should they do with it?
They might spend more time helping a new employee understand the organization's culture. They might identify barriers that an automated workflow cannot recognize. They might coach managers through difficult situations. They might notice that an employee's experience is deteriorating before a survey captures it. They might solve the exception rather than processing the routine.
That is the point of automation. The objective is not to remove humanity from work. It is to remove unnecessary work from humans.
This distinction becomes even more important as organizations redesign jobs around AI agents. Microsoft's 2026 Work Trend Index describes an emerging shift toward organizations in which AI agents increasingly execute tasks and workflows while people provide direction, judgment, and oversight. The implication is significant: as machines take on more execution, organizations can no longer define human value primarily by how much activity an employee completes. They must become better at defining the quality of decisions, relationships, and outcomes employees create.
A second leadership challenge hides inside this transition.
Organizations cannot simply automate every repetitive task and assume the result will automatically be better. Some routine work is also developmental work. People often acquire professional judgment by first handling simpler responsibilities and gradually confronting more complex situations. If AI eliminates those experiences without creating alternative learning pathways, organizations may eventually discover they have automated the very activities that develop future experts.
The answer is not to preserve inefficient work for tradition's sake. It is to redesign development alongside the work itself.
If AI produces the first analysis, employees need to learn to interrogate it. If an agent handles routine customer interactions, employees need greater exposure to complex customer situations. If automation eliminates administrative responsibilities, organizations need to deliberately create opportunities for employees to develop strategic, interpersonal, and decision-making capabilities.
This is the difference between workforce reduction and workforce transformation. One treats recovered capacity as a cost-saving opportunity. The other treats it as organizational capacity that can be redeployed toward higher-value work. That distinction will increasingly determine whether AI creates durable competitive advantage or simply produces another generation of productivity tools.
It also changes what executives should measure.
Counting AI licenses, training hours, chatbot interactions, or pilot programs can tell leadership whether the organization has adopted the technology. Those metrics cannot tell leadership whether the organization has changed.
The more meaningful measures are operational. Has cycle time declined? Have unnecessary handoffs disappeared? Has error frequency decreased? Has employee or customer experience improved? Are experts spending more time on consequential decisions? Has the organization increased its capacity without simply increasing the amount of work people are expected to absorb?
These questions force AI strategy back into business language.
They also expose an uncomfortable truth: many organizations do not have an AI problem. They have a process problem that AI is making impossible to ignore. Leaders who understand this will approach AI differently. They will not begin by asking which tool their organization should purchase. They will begin by examining how work actually moves through the organization, identifying where human judgment creates value, and challenging every activity that survives primarily because it has become familiar.
That requires courage because processes are rarely emotionally neutral. They determine roles, budgets, authority, and status. When a leader asks why a human is still performing a particular task, the answer may reveal that an entire position, department, or management assumption was built around work that no longer needs to be performed manually.
The technology decision is often the easy part. The harder decision is what happens to the human capacity that technology releases. That is where responsible leadership begins.
The future will not belong to organizations that simply have more AI. It will belong to organizations that understand what humans should stop doing, what humans should become exceptionally good at, and where the boundary between the two should keep moving.
So the next time an AI initiative is presented as transformation, look past the platform, the dashboard, and the demonstration. Look at the work. Find a process that consumes significant human attention. Follow it from beginning to end. Examine every approval, handoff, reconciliation, notification, and decision.
Then ask the question most organizations still avoid: Why is a human still doing this?
If the answer is that the work requires judgment, empathy, context, or accountability, protect that human contribution. If the answer is simply that the organization has always done it that way, redesign it.
That is not merely an AI strategy. It is a leadership discipline for deciding what human effort is actually worth. And organizations that learn that discipline will do more than deploy AI. They will build organizations worthy of the humans who lead them.
Live, Lead & Leave A Great Legacy! ~ Izabela Lundberg
Related: AI Is Entering Its Most Dangerous Phase for Investors