The Josh Bersin Company has released a new HR 2030 research guide outlining how organizations can manage growing numbers of AI agents without creating fragmented or competing systems. The report, The HR Leader’s Guide to Multiagent HR Architecture, recommends organizing specialized HR agents into coordinated functional groups, supported by shared data, governance, monitoring and human oversight.
The rapid adoption of AI agents is creating a new challenge for HR technology leaders: managing too many autonomous systems rather than simply finding more ways to deploy them.
The Josh Bersin Company is addressing that challenge in its latest HR 2030 research, The HR Leader’s Guide to Multiagent HR Architecture. The report provides CHROs and CIOs with a framework for coordinating specialized AI agents while controlling data, governance and operational risks.
The research identifies at least 151 AI agents currently attracting interest from CHROs. Rather than treating each agent as an isolated automation tool, the report argues that organizations need an architecture capable of coordinating groups of agents around broader HR functions.
The concept is built around what the report calls “superagents”: functional families of specialized AI agents that work together under a common architecture. Individual agents can focus on specific activities while the broader family coordinates their actions, data and objectives.
That distinction could become increasingly important as enterprises move from generative AI assistants toward agentic systems capable of executing multi-step workflows.
Recruiting illustrates the problem. The Josh Bersin Company describes an East Coast bank that had developed 17 internal agents to support activities including sourcing, candidate selection, interviewing and interview scheduling. According to the research, those agents were not coordinating with one another or consistently accessing the same data, producing results that became inconsistent and less trustworthy.
The example highlights a potential weakness in agent-by-agent deployment. Automating individual tasks can create efficiency within a workflow while simultaneously introducing fragmentation across the broader HR operating model.
A multiagent architecture attempts to address that problem by establishing coordination at a higher level.
The research recommends organizing agents into eight core functional HR families covering areas such as hiring, workforce management, compensation, retention and employee mobility. Specialized agents can remain responsible for individual tasks while operating within a larger structure designed to maintain consistency and shared context.
The report estimates that this model could eliminate up to 40% of current task-oriented HR work. That figure is a projection from the Josh Bersin Company rather than an independently validated benchmark, but it illustrates the scale of automation the firm believes coordinated agent architectures could enable.
The technology foundation extends beyond the agents themselves. The report emphasizes data governance, orchestration and context layers, a data fabric, monitoring systems and a shared skills library.
These components become particularly important when AI systems move from generating recommendations to taking actions. An agent that can initiate recruiting activities, access employee information or make decisions within a workflow requires clearly defined permissions and monitoring.
For HR leaders, that makes agent architecture a governance issue as much as a technology issue. The report argues that decisions about how agents interact, what data they can access and where humans remain involved should be treated as business design decisions.
The approach also separates organizational AI agents from personal AI tools. The research cites tools such as Muse, GrokBot, Dots and Autopilot as examples of personal agents that can support individual productivity and employee experiences without necessarily becoming part of the organization’s core HR operating architecture.
This distinction could become increasingly relevant as employees bring their own AI assistants into workplaces while enterprises simultaneously deploy centrally governed agents.
The resulting HR technology environment may therefore contain several layers: personal AI tools used by individuals, specialized enterprise agents performing specific HR tasks, and higher-level agent families coordinating those systems.
Monitoring becomes another critical component. AI agents can produce unpredictable outputs or operate outside intended boundaries, creating risks around employee data, compliance, decision-making and accountability. Continuous monitoring and optimization can provide organizations with mechanisms to identify and correct those issues.
The Josh Bersin Company argues that the objective should not be to determine which individual AI agents an organization should purchase. Instead, HR and technology leaders need to design how multiple agents operate together as part of a coherent workforce architecture.
That represents a significant change in the way organizations may evaluate HR software. Traditional application decisions often focus on features, integrations and user experience. Agentic HR introduces additional questions around orchestration, autonomy, data access, model behavior and human intervention.
For CHROs and CIOs, the emerging priority may therefore shift from AI adoption to AI architecture. The organizations most likely to capture value from HR agents may be those that establish clear boundaries and coordination mechanisms before agent deployments proliferate.
The research is the second release in the HR 2030 initiative. Its next phase is expected to examine how a multiagent operating model could affect individual HR workflows and the roles of HR professionals.
Market Landscape
Enterprise HR technology is moving from isolated AI copilots toward agentic systems capable of executing multi-step workflows. As vendors and enterprises introduce specialized agents for recruiting, employee support, workforce planning and other functions, the risk of fragmented deployments is increasing.
The multiagent architecture described by the Josh Bersin Company points toward an emerging AI orchestration layer for HR. Shared data, skills libraries, governance, monitoring and human oversight could become important components of enterprise HR architecture as autonomous systems take on more operational responsibilities.
For HR technology buyers, this means evaluating not only what an AI agent can do, but also how it interacts with other agents, enterprise data and human decision-makers.
Top Insights
- Josh Bersin Company identifies at least 151 AI agents attracting interest from CHROs, increasing the need for coordinated enterprise architecture.
- The report recommends grouping specialized agents into functional HR families or “superagents” instead of deploying isolated automation tools.
- Shared data, orchestration, monitoring, governance and skills libraries are positioned as foundational components of multiagent HR.
- The research estimates coordinated agent architectures could eliminate up to 40% of task-oriented HR work.
- Human oversight remains central as organizations manage agent autonomy, unpredictable outputs, data access and operational risk.
Join thousands of HR leaders who rely on HRTechEdge for the latest in workforce technology, AI-driven HR solutions, and strategic insights
