your next direct report might NOT be human
By Sonia Nash
AI agents are arriving into the workforce and most organisations see them as tools that sit inside softwares. But AI agents have started to become active contributors and gain more autonomy to write, delete and execute within organisations’ systems. They work alongside humans, hand off tasks, and in some cases, they manage end-to-end workflows. And yet the function responsible for managing the workforce - Human Resources - was built entirely around the assumption that all workers are human.
The assumption no longer holds.
In 2022, as I watched a video by OpenAI on the rise of AI Agents, one thought that crossed my mind then is that HR as we know it will need to evolve. My personal view on the matter is that HR should be renamed as WM: Workforce Management. Not the term that is used in some manufacturing environment, but as reimagined strategic function - one that governs a workforce made of humans AND AI agents.
But it’s more than a name change:
Managing a Blended Squad
The next evolution is already underway: managing a blended squad composed of humans paired with AI agents, working together towards a shared goal/outcome. This brings about a new set of skills and capabilities within WM - deciding which roles and tasks belong to human and which ones get delegated to agents. The organisations that will win won’t be the ones with most agents, or with most humans. It will be the ones who know how to combine them.
WM leaders will need a new ability: that of looking at a business function and switch from “how many people do we need” to “what’s the right human-agent combo for this work and how do we make this team perform”.
Agents have a lifecycle too - and someone needs to own it
HR currently owns the employee lifecycle. From attraction and onboarding through development, performance management and eventually offboarding, the journey of a human contributor is HR's domain. So how do we factor in AI agents?
Agents also have a lifecycle: They need to be onboarded — configured, scoped, given the right context and constraints. They need to be performance-managed — are they producing what was intended? Are outputs drifting over time? Are they operating within ethical guardrails? And they need to be offboarded — deprecated when better models emerge, replaced when requirements change, or retrained when business needs shift.
Right now, this work sits awkwardly between IT, operations, or no one in particular. That's a governance gap, and it carries real risk. WM will close it. New roles will need to emerge of course — Digital Workforce Leads, perhaps — professionals who sit at the intersection of HR, technology, and business operations, and who are accountable for the performance and conduct of the AI agents in their care.
Culture and the human experience cannot be relegated
One of HR’s most important function is creating and maintaining the conditions for human work: psychological safety, belonging, a sense of meaning and contribution. Employees need to be valued and feel valued. Introducing AI agents into the daily working experience can create genuine fear and friction for many people. Am I being replaced? Is my performance being monitored by an AI I can’t see? Who gets credit when a human-agent team does great work? How does that affect my compensation and opportunities for career progression?
These are not irrational fears and they deserve honest answers through a function that owns that narrative. WM will need to set clear guidelines around how agents are used, what humans remain accountable for, and how contribution is recognised in a blended environment. The human experience cannot become an afterthought in the drive towards efficiency. If people feel like they are competing with their digital colleagues rather than collaborating with them, the productivity gains of AI will be swallowed by disengagement and attrition.
What human skills in a blended squad?
Learning and development, another core HR process, has always been about closing the gap between the skills people have and the skills the business needs. In a world of AI agents, that gap changes shape entirely; it is not only between human skills and business requirements but also between humans and the ever-expanding capabilities of AI agents.
Now it’s not all doom and gloom; as AI agents absorb more of the transactional and analytical load, the skills that remain distinctly human will become more valuable. The ability the critically evaluate agent output; the judgement to know when to override a model’s recommendation; the guts to ask a difficult question; the emotional intelligence to navigate the human moments, like hearing what was not said in that meeting or reading between the lines and acting accordingly.
Along with soft-skills, a new form of operational literacy will become essential: the ability to work with agents effectively. That means, knowing how to prompt, how to interpret agent output and catch anomalies, and how to integrate an AI agent’s output into a blended piece of work. A lot of these skills currently sit with engineers or data scientists, but they will need to become skills of any business functions working with AI agents.
Workforce risk now includes AI agent risk
When something goes wrong with a human employee, HR deals with it. The accountability is very clear. But in a blended squad, the questions of accountability becomes a lot more complicated. When an AI agent operating with significant autonomy makes a decision that harms the business, if an AI agent handles a customer interaction in a way that breaches policy, where does the accountability land?
HR has historically owned the people-related risks: there’s employment law of course, DEI compliance, codes of conduct etc. If we assume HR transforms into WR, we imply that workforce risk - read here the entire workforce (human and AI agents alike) - is owned by WR.
That will require WM leaders to develop literacy in AI governance to understand how models are audited, how bias is identified and corrected, how AI agent behaviour is logged and reviewed, and how/where accountability is assigned when something goes wrong. And that shouldn’t sit with IT alone.
Where to next?
The organisations that treat AI agents as software to be managed by IT will lose to the companies who treat them as workforce members to be managed strategically.
AI agent may not be employees in any legal or even philosophical sense, but their impact on how work gets done, how human feel about their work and their value and how organisations perform is workforce-level question. And it needs a function with the capability and mandate to answer it.
HR has reinvented itself before, from personnel administrators to strategic business partners. It can do it again. However, the workforce landscape has never been more complex. The question for every HR leader reading this is simple: are you ready to manage a workforce that isn’t entirely human? Because ready or not, that workforce is already here.