Companies are bringing people strategy and technology delivery closer as artificial intelligence changes jobs, skills and operating processes. The harder question is not whether the functions should cooperate, but whether one executive, two coordinated leaders or a temporary fusion team offers the clearest accountability.

On 15 September 2025, Harvard Business Review published a debate by Gretchen Gavett and Thomas Stackpole drawing on Keith Ferrazzi, Peter Cappelli and Kalifa Oliver. Their recommendations differ. The article is therefore more useful as a map of organizational choices than as an endorsement of a combined chief human-resources and technology post.

Three models solve different coordination problems

A merged role puts decisions about systems, skills and workforce change under one executive. Co-leadership keeps specialist authority separate but binds the leaders to shared outcomes and a common transformation agenda. A fusion team assembles the required functions around a defined use case, with authority returning to the permanent organization after delivery.

These structures are not stages in a maturity ladder. A company may use more than one at once: a shared executive for enterprise platforms, two leaders for workforce planning and a temporary team for an AI-enabled service. Titles also vary between companies, so the useful comparison concerns decision rights, expertise and capacity rather than labels.

Three separate architectural structures show a bridge joining people and digital systems, two independent towers connected by one rail, and a compact team assembled from interlocking modules
The organizational choice should follow the coordination problem instead of a fashionable title.

Eight questions before changing the chart

  1. What recurring decision or delivery failure is the redesign meant to correct?
  2. Which measurable business and workforce outcomes should improve?
  3. Does one candidate possess enough depth across both domains?
  4. Will the combined remit leave sufficient attention and operating bandwidth?
  5. Which decisions require one accountable owner, and which benefit from challenge?
  6. Can shared data, definitions and access controls support joint work?
  7. How will legal, financial, security and operational expertise enter the decision?
  8. What evidence and review date will determine whether the model should continue?

A merged role can remove a handoff and create a bottleneck

Moderna is cited in the debate as an example of combining technology and human-resources leadership. The logic is direct: AI deployment changes workflows and capability at the same time as it changes software and data. One accountable executive can settle priorities across both domains and prevent a platform programme from arriving without job redesign or training.

The trade-off is concentration. Human resources covers employment practice, compensation, talent and organizational health; technology leadership covers architecture, resilience, security, data and delivery. A larger title does not create expertise or hours. The model is credible only when strong deputies retain domain authority and the executive has a clear, limited mandate.

Co-leadership preserves specialist challenge

Gap is presented as a case in which people and technology leaders co-created a transformation “North Star” without making a formal merger the necessary first move. That arrangement can preserve distinct professional judgment while forcing agreement on outcomes, sequencing and investment.

Co-leadership fails when collaboration depends on goodwill alone. It needs a written decision map: who recommends, who approves, how disagreement is resolved and which measures both leaders own. A joint steering meeting is not a substitute for those rules. Nor does a shared presentation prove that frontline work, incentives or systems have changed.

For companies operating in the United States, the design also has to accommodate employment, privacy, security and sector obligations that may assign responsibilities to different officers. Governance should make these interfaces explicit instead of assuming an unusual title overrides them.

Fusion teams work when the boundary is a use case

Cisco is cited for an AI Control Hub that spans human resources, technology, legal, finance and operations. This illustrates a third approach: keep functional leadership intact while creating a cross-functional environment for decisions that no single department can make safely. The team can define permitted data, evaluation, funding, operational ownership and escalation for a portfolio of uses.

A fusion team should have an entry condition, deliverables and an exit route. Otherwise it becomes another approval layer. Each initiative needs a named business owner, a technical owner, affected-workforce representation, risk classification, evidence threshold and stop authority. Permanent functions remain responsible for standards and professional judgment.

Five distinct miniature workstations for workforce capability, computing, compliance, finance and operations exchange transparent evidence capsules through illuminated approval gates
A shared control environment adds value when evidence moves with the decision and accountability remains visible.

Evidence that the structure is working

  • time from an approved use case to controlled deployment;
  • rework caused by late workforce, security, legal or data review;
  • adoption and competent use by the employees whose jobs change;
  • incidents, near misses and interventions by risk class;
  • operational outcomes compared with the previous process;
  • decision delays and workload carried by each accountable leader.

Start with the problem, then test the operating model

A Talent Edge Weekly summary highlights the underlying disagreement among the experts and the need to examine context. That disagreement is productive: structural simplicity, specialist depth and distributed participation cannot all be maximized by the same design.

Management should state the failure it is trying to fix, choose the smallest structural intervention that can fix it and set a review date. A reporting-line change is an input, not evidence of AI adoption. Evidence appears when decisions become faster without hiding risk, employees can use the new process, and business results improve under defined controls.