Published
Author
There is much discussion about generative AI in HR right now. The ability to quickly create content, formulate goals, or summarise information has already made a tangible difference in everyday work. But the real shift will come when AI stops being reactive support and instead becomes an active part of how the organisation governs, follow up, and act on talent and workforce capability.
This is where agentic AI comes in – especially at the intersection of learning, skills and talent.
In many organisations, these areas are still managed as separate tracks: learning resides in an LMS, skills are defined in frameworks and taxonomies, and recruitment or mobility are handled in separate processes. If this sounds familiar, you are not alone. My thesis is simple: learning, skills and talent are three perspectives on the same underlying capability layer, powered by a shared skills logic. When they are not connected, both investments and efforts become fragmented. Efforts become reactive, investments have less impact, and it takes longer to build the skills the business actually needs.
With agentic AI, the conditions change.
AI has quickly become an important support in HR and talent management. So far, much of the focus has been on generative AI – a technology primarily used to streamline individual tasks, such as developing learning plans, formulating goals, summarising information, or creating job descriptions. It delivers productivity gains but still requires someone to initiate the need.
With agentic AI, a shift occurs. We move from AI that responds to prompts to AI systems that can interpret goals, plan, and act over time within defined parameters. It is no longer just about recommending, but about initiating, coordinating, and following up on activities related to learning, skills, and talent.
This is why agentic AI becomes particularly powerful at the intersection of these areas. Where generative AI improves parts of existing processes, agentic AI creates the conditions for a more connected and continuous approach. Learning can be triggered when needs arise, skills development becomes embedded in real work, and organisations can act more proactively in talent decisions.
Rather than seeing agentic AI as yet another tool for L&D, recruitment, or talent management, it is more useful to view it as an orchestration layer – a system of action that connects these domains. The boundaries between disciplines begin to blur, and the focus shifts from optimising individual processes to strengthening the organisation’s ability to identify, develop, mobilise, and apply skills.
Agentic AI refers to systems that work towards goals – not just prompts – and that often consist of multiple specialised agents collaborating in orchestrated flows, with human oversight and clear governance. In an HR context, this means that AI can:
In a learning–skills–talent context, agentic AI can continuously work towards goals such as reducing the time from change to applied skills, increasing internal mobility, improving matching between people and opportunities, and maintaining critical skills as work evolves.
As AI-driven flows become more autonomous, transparency, traceability, and accountability become increasingly important. Organisations need to understand why recommendations are made, which data is used, and who is ultimately responsible for decisions affecting people and careers.
When discussing agentic AI, it is easy to imagine it as a new module in an LMS, ATS, or HCM system. In practice, that view is too narrow. What is emerging instead is an ecosystem of specialised AI agents collaborating around a shared skills architecture and common goals. For example:
The key point is that none of these agents create value in isolation. Value emerges when learning, skills, and talent are connected in orchestrated flows – not when individual processes are optimised separately.
Recruitment is an area where agentic AI is already becoming tangible. Where AI previously focused on matching CVs to static requirements, it is now evolving towards understanding roles in context and inferring skills, adjacent capabilities, and potential.
This means that skill requirements can be dynamically interpreted based on real work, role definitions can evolve as needs change, and outcomes from hiring can be fed back into the organisation’s skills architecture.
When the same skills logic is used across recruitment, internal mobility, and learning, a connected talent ecosystem begins to emerge. Agentic AI can make that connection operational – not as a PowerPoint ambition, but as a true system of action embedded in everyday decisions.
An often underestimated challenge is that the difficult part is rarely the AI technology itself. The real challenge is organising skills data so that it is consistent and meaningful across the entire talent ecosystem. In many organisations, skills data is spread across multiple systems – learning platforms, HR systems, recruitment tools, and workforce solutions. The issue is rarely a lack of data.
The issue is that the same skill can mean different things in different contexts. When that happens, it becomes impossible to use skills consistently for learning, matching, and decision-making. For agentic AI to work effectively, a shared skills architecture and common semantics are essential.
Many organisations try to solve fragmentation by adding more integrations between systems. But the bottleneck is rarely technical. More often, it is the lack of a shared understanding of skills, how they develop, and how they should be applied in workforce decisions. In an agentic context, skills data needs to function as governing information. It must be dynamic, continuously updated, and linked to actual performance – not confined to a single system. Otherwise, there is a risk that inconsistencies are simply automated rather than resolved.
There is also reason for caution. Many organisations are still at an early stage when it comes to skills architecture, data quality, and shared definitions.
If different systems rely on different definitions of skills, performance, or potential, agentic flows risk reinforcing existing inconsistencies rather than solving them. At worst, ambiguity becomes automated.
Agentic AI acts as a maturity indicator for how well an organisation has established shared definitions, data quality, and governance around skills. Moving from pilots to real impact requires several shifts:
Agentic AI therefore requires not just technology, but leadership.
As AI takes on a more active role, HR’s mission evolves. HR needs to work more data-driven, act as a steward of skills data, and be able to interpret and quality-assure AI-driven decisions. At the same time, the role becomes more strategic – orchestrating the interaction between people, business needs, and digital capabilities. The result is a more continuous and business-relevant approach to workforce capability.
Agentic AI makes it increasingly difficult to manage learning, skills, and talent as separate tracks. Instead, they form a connected capability system that must be led as a whole. This is not just a technical shift – it is an organisational transformation.
The organisations that succeed will not be those that implement the most AI features, but those that manage to unify learning, skills, and talent into a connected ecosystem where people, business, and AI work towards shared goals. Ultimately, this creates more contextual and continuous support for both managers and employees.
The key question is therefore not how autonomous AI can become, but whether the organisation has the capability, governance, and data foundation required to use that autonomy responsibly.
Many organisations are facing the same questions: how to connect HR systems, how to establish a shared skills architecture, and how to move from initiatives to measurable impact.
This is where external expertise can make a meaningful difference.
