A leadership team approves an AI tool, then asks HR a deceptively simple question: “Which roles can it replace?” That question sets the organization on the wrong course before implementation has begun. The stronger question is how to redesign AI roles so the technology absorbs low-value work while experienced people take on work that demands judgment, trust, context, and accountability.
AI adoption is not primarily a headcount exercise. It is a work redesign exercise. Organizations that treat it otherwise may achieve a short-term productivity gain while quietly eroding the institutional knowledge, decision quality, and leadership bench they will need to operate through the next disruption.
For CHROs and senior leaders, this is a strategic opening. HR can move the conversation beyond job elimination and toward a more valuable agenda: building an organization where human capability and machine capability reinforce one another.
Why redesigning AI roles is an executive decision
A job title is not a unit of work. It is a bundle of tasks, relationships, decisions, risks, and accountabilities assembled over time. AI can perform some tasks within that bundle exceptionally well. It can summarize large volumes of information, draft routine communications, identify patterns, generate first-pass analyses, and reduce administrative drag.
It cannot own the consequences of a sensitive employee relations decision. It cannot reliably interpret the political, cultural, or customer context behind a performance issue. It cannot earn credibility with a skeptical executive team, coach a manager through conflict, or determine when a statistically sound recommendation is ethically wrong for the organization.
That distinction matters because poorly designed AI adoption often strips roles of the developmental work that produces future leaders. If junior analysts no longer synthesize information, how will they learn to challenge assumptions? If managers outsource difficult conversations to generated scripts, how will they build the confidence to lead people? If experienced employees are treated as expensive legacy capacity rather than essential interpreters of organizational context, the company loses its operating memory at the exact moment change requires it most.
The goal is not to preserve every task. The goal is to protect and expand the work that creates enterprise value.
A practical framework to redesign AI roles
The most useful starting point is not the org chart. Start with the work. Review roles through four lenses: automate, augment, elevate, and govern.
Automate work that adds little human value
Automation is appropriate when work is repetitive, rules-based, high-volume, and low-consequence if reviewed. Scheduling, data entry, document formatting, standard reporting, basic knowledge retrieval, and first-draft communications often belong here.
Removing this work can create real capacity, but leaders should be precise about what happens to that capacity. “More strategic work” is not a workforce plan. Define the higher-value activities the role will now own, the skills required, and the measures that will show whether the shift is producing value.
Augment work where speed improves quality
Many professional roles will not be replaced by AI. They will be reconfigured around AI-supported execution. A recruiter can use AI to prepare interview materials and synthesize candidate themes, then spend more time assessing motivation, culture contribution, and hiring-manager alignment. An HR business partner can use AI to surface workforce patterns, then focus on the conversations and interventions those patterns require.
Augmentation works when the human remains actively engaged in the reasoning process. The risk is passive acceptance. When people treat AI output as an answer rather than an input, speed can become a substitute for judgment.
Elevate work that requires human judgment
This is where redesign creates strategic value. As AI reduces transactional workload, roles should move upward into problem framing, decision support, relationship stewardship, change leadership, and capability building.
For HR leaders, this may mean redesigning a people manager’s role to include more coaching, team effectiveness, workforce planning, and risk sensing. It may mean shifting HR operations professionals toward process ownership, data quality, employee experience design, and vendor governance. It may mean giving experienced employees formal responsibility for mentoring colleagues who are learning to work with AI.
Elevation is not simply adding more work to a role. It is deliberately changing the level of contribution expected from that role.
Govern work where risk and accountability remain human
Every AI-enabled role needs clear boundaries. Who validates output? Which decisions require escalation? What data can be used? When must employees disclose that AI was involved? What happens when a recommendation conflicts with policy, law, or professional judgment?
Governance cannot sit solely with legal, IT, or an AI council. It has to be built into job design, workflows, manager expectations, and performance conversations. The employee using the tool should understand where authority ends and accountability begins.
Do not confuse task reduction with role simplification
The temptation in a cost-constrained environment is to assume that fewer tasks mean a simpler, lower-level job. Sometimes that will be true. More often, the opposite is true.
When routine work disappears, what remains may require greater discernment. Consider a customer service role where AI handles straightforward inquiries. The human cases are now the escalations: angry customers, unusual circumstances, exceptions, and moments where retention is at risk. The role has fewer interactions but higher emotional intensity, commercial consequence, and decision complexity.
This has implications for job architecture, compensation, staffing levels, and manager capability. A redesigned role may require fewer people, more skilled people, or a different mix of people. It depends on the work volume, risk profile, customer expectations, and quality standard. Executive teams need to resist broad assumptions about efficiency until they have examined those variables.
Make experienced talent part of the design process
Organizations often make a costly mistake by positioning experienced employees as the population most threatened by AI. Experience can be one of the strongest assets in role redesign because experienced people recognize the exceptions that are absent from process maps and training data.
They know which customer commitments create downstream problems, which policy language managers misinterpret, which workarounds keep operations moving, and which decisions look easy until conditions change. That knowledge is not nostalgia. It is operating intelligence.
Invite experienced employees into role-mapping sessions, pilot reviews, and workflow redesign. Ask them where automation could create hidden risk, where it would remove meaningful friction, and what newer employees will need to learn if AI takes over foundational tasks. Their participation also increases adoption because people are more likely to trust a new workflow when they can see how practical expertise shaped it.
This does not mean protecting outdated work merely because it is familiar. It means distinguishing between work that is obsolete and knowledge that remains strategically useful.
Build capability before demanding adoption
AI fluency is becoming part of workforce capability, but fluency is more than prompt writing. Employees need to know how to assess source quality, challenge flawed outputs, protect confidential information, recognize bias, and decide when not to use the tool.
Leaders need a different layer of capability. They must be able to set clear outcomes, redesign team workflows, spot uneven adoption, manage workload shifts, and hold people accountable for the quality of AI-assisted work. A manager who cannot explain how a role has changed will struggle to lead the person in it.
A practical approach is to define a small set of role-specific behaviors. For example, an HR leader might be expected to use approved AI tools for analysis and drafting, validate outputs against organizational context, document material decisions, and escalate high-risk use cases. These expectations are more useful than generic encouragement to “embrace AI.”
Measure whether redesign is working
Productivity matters, but it is not enough. A role redesign effort should be measured across operational, people, and risk outcomes.
Operational measures may include cycle time, error rates, service levels, and quality of decisions. People measures may include skill growth, manager confidence, workload sustainability, internal mobility, and employee sentiment. Risk measures should track data handling, policy exceptions, bias concerns, and instances where AI output required correction or escalation.
If output increases while quality, trust, or employee capability declines, the organization has not modernized. It has simply moved cost and risk to another part of the system.
The leadership question that matters
The organizations that gain the most from AI will not be those that adopt tools fastest. They will be those that redesign work with the clearest view of what humans are uniquely positioned to do.
For CHROs, this is a moment to lead from the center of business strategy. Bring the work, skills, accountability, and experience conversation to the executive table before technology decisions harden into workforce decisions. The future of work will be shaped role by role, and the choices made now will determine whether AI reduces capability or compounds it.
Bring Lorraine to your team.
Speaking engagements, executive workshops, and HR leadership strategy.