Executive summary
Empathy, judgment, and meaning create value when people have a clear purpose, trusted context, and authority to act. Leaders must design those conditions into AI-enabled work and measure the contribution they produce.
When an AI system can prepare the analysis and draft the response, the human role needs a clearer purpose. A customer may need someone to understand an unusual circumstance. A team may need a decision between competing priorities. An executive may need to explain which outcome matters and who will answer for it.
The World Economic Forum’s Future of Jobs Report 2025 reflects that continuing demand. Surveyed employers ranked analytical thinking as their leading core skill, with leadership and social influence, empathy and active listening also among the top ten. These are employer assessments of workforce needs; they do not establish that any capability is permanently beyond AI. [1]
The operating signal is that automating parts of a job leaves a design decision about the human contribution. Its value depends on the responsibility people carry and the conditions in which they exercise it.
The human edge becomes useful when empathy, judgment, and meaning have a defined place in the work. Empathy helps people uncover circumstances that a standard record may miss. Judgment connects evidence with priorities and consequences. Meaning gives people a clear reason to care about the result and carry responsibility through to completion.
AI can support those activities. The organizational question is who is accountable for the decision, the relationship, and what happens next. That accountability needs competent people, sufficient context, and practical authority.
If a redesigned role consists mainly of approving recommendations under pressure to move faster, the organization may weaken the very contribution it intends to preserve. Calling someone a reviewer does not establish whether they can recognize a problem, challenge it, or change the outcome.
The operating idea
Understand the person.
Surface missing context.
Weigh the tradeoff.
Challenge the recommendation.
Connect work to purpose.
Own the outcome.
Leadership owns
Clear responsibilities · Authority to act · Trusted context · Time to think
↶ Experience and feedback improve the next decision
For each role affected by automation, name the decisions, relationships, and outcomes it owns. Be specific: resolving an unusual customer situation, challenging an unsupported recommendation, or deciding when an exception exposes a policy problem. Then identify the knowledge and experience required to do that work well.
Keep routine work automated where performance and risk justify it. Concentrate human attention where context, consequences, or competing interests require a deliberate decision.
Specify when a case reaches a person, what evidence arrives with it, what that person may change, and where they can obtain help. Give them a usable route to pause an action or escalate it before the consequence occurs.
NIST’s AI Risk Management Framework calls for clear responsibilities, training, and defined roles in human–AI arrangements. It also calls for oversight processes to be assessed and documented. The practical implication is to test whether people can exercise the responsibility assigned to them. [2]
In practice / Illustrative example
An AI assistant recommends rejecting a refund because the request falls outside the standard window. The service employee discovers that the delay followed repeated failures in the company’s own returns process. Their role is to establish the facts, apply an authorized exception or escalate it, explain the resolution, and flag the recurring process defect.
The human contribution is visible in the decision and the follow-through. The workflow must provide case history, exception authority, and time to investigate. Measuring only handling speed would miss much of that value.
Where automation releases usable time, decide how much goes to additional service, difficult cases, coaching, improvement, or cost reduction. Put that choice into staffing and workload plans. A promise of more meaningful work needs actual assignments, capacity, and recognition.
Retain opportunities to practice judgment. Reviewing a difficult case with an experienced colleague can make missing evidence, competing priorities, and the reasons for a decision visible.
Review repeat contacts, reopened cases, resolution quality, and recurring exceptions alongside speed and cost. Examine why people accepted, changed, or escalated recommendations. An override count alone cannot distinguish sound judgment from unnecessary intervention.
Give an owner responsibility for turning those findings into better guidance, policies, or system behavior. Human experience should improve the operating system over time.
Before approving the benefits case for an automated workflow, leaders should be able to describe the roles that will operate it: their purpose, decision authority, capability requirements, and realistic workload. The Human-in-the-Loop Synergy framework develops that design question in more depth.
Be explicit about the workforce choices and tradeoffs. People need an honest account of how their work will change, what they are expected to contribute, and what support will be available.
The human edge requires deliberate investment. Leaders make it credible through the work they assign, the authority they grant, and the outcomes they recognize.
For one role being redesigned around AI, what consequential decision will its people own—and what in the workflow gives them the context, authority, and time to make it well?
Short, practical perspectives on AI-era operations, governance, and operating-model transformation.
Subscribe
RePerspective Labs