The End of the Three-Pillar HR Model

What structurally obsolete means — and what actually changes in roles, teams, and governance

What Is Actually Ending

The three-pillar HR model is not under pressure. It is structurally obsolete — and the distinction matters for how HR leaders respond.

Under pressure implies the existing model needs reinforcing, accelerating, or protecting. Structurally obsolete means the logic on which the model was built no longer holds. Those are different diagnoses. They produce different responses. And most HR functions are currently applying the first response to the second problem.

Dave Ulrich's three-pillar model — HR Business Partners embedded in the business, Centres of Excellence providing specialist capability, Shared Services handling transactional volume at scale — was not an arbitrary design choice. It was a logical response to three structural conditions: HR held information that others lacked, the volume of HR transactions required dedicated processing capacity, and some HR work was complex enough to require specialist synthesis. Each pillar existed because a structural condition justified it.

Agentic AI is dismantling all three conditions simultaneously. That is what makes this moment different from every previous technology wave in HR.

Why Simultaneous Erosion Is the Critical Point

Previous HR technology waves were additive. They improved the efficiency of the existing model without challenging the rationale for its structure. HRIS platforms reduced manual data entry. Workflow automation accelerated transaction processing. First-generation chatbots handled simple queries. In each case, the three structural conditions remained intact.

Agentic AI is different because it erodes all three conditions at once.

An AI agent operating on an enterprise knowledge base does not merely process faster — it closes information asymmetry in real time. The HRBP's edge in navigating complex employment law situations narrows when an agent can answer the same question in seconds with jurisdiction-specific accuracy.

An AI agent designed for high-volume transactional workflows does not merely support Shared Services — it removes the structural rationale for it. When an agent completes onboarding, answers payroll queries, manages leave requests, and handles benefits administration end-to-end, the volume that justified a dedicated function no longer justifies its current size or shape.

An AI agent capable of analysing engagement data across thousands of employees, modelling workforce scenarios, and generating strategic recommendations does not merely assist the CoE — it changes what the CoE professional is for. The value shifts from producing the analysis to interrogating, challenging, and governing its use.

Three conditions gone — not sequentially, not partially, but simultaneously. TI People's AI-Powered HR research is explicit: the three-pillar model will not survive AI intact. Significant rearchitecting is required. Not an incremental adjustment. Rearchitecting.

What Actually Changes — and When

Re-architecting is already underway in organisations paying attention. Here is what it looks like in practical terms over the next few years.

Shared Services contracts rather than disappear. The transactional volume that justified large Shared Services functions does not vanish — it migrates. AI agents handle the routine end-to-end. What remains is the exception layer: complex queries, edge cases, complaints, and the governance of the agents themselves. The Shared Services professional of 2027 is not processing transactions. They are monitoring AI outputs, managing escalations, and maintaining the systems that handle the volume. Fewer people. Different skills. Different accountability.

The HRBP role polarises. Business Partners in the middle of the value spectrum face the greatest exposure. Their value has historically rested on serving as the conduit between business and HR expertise. When AI closes the information gap and handles much of the routine advisory work, the conduit role thins. What survives at either end: the deeply relational HRBP who is trusted by business leaders for their judgement, and the technically capable HRBP who can design and govern AI-augmented people processes. The middle is where the pressure lands hardest.

Centres of Excellence shift from production to governance. The CoE has historically produced the frameworks, policies, and analytics the business needs. When AI can produce the first draft of a competency framework, the initial engagement data analysis, or the scenario modelling for workforce planning, the CoE's role shifts from production to interpretation, governance, and quality assurance of AI outputs. The premium moves from the analyst who builds the model to the expert who challenges its assumptions.

New roles emerge that do not yet have names. The organisations furthest ahead are already creating roles that sit outside the three-pillar architecture entirely — AI systems governance lead (owning conformity, auditability, and failure modes), workforce intelligence interpreter, human-machine workflow designer. These are not HRBP variants or CoE analysts with different titles. They are a new category of HR professional defined by their relationship to AI systems rather than by their position in a functional hierarchy.

The Question This Puts to Every HR Leader

The structural transition described above is already visible in organisations that are designing for it — and is about to become visible, under duress, in the organisations that are not.

The question it puts to every HR leader is precise: are you redesigning your operating model around what AI makes possible, or are you optimising the existing model for a world that is already changing beneath it?

Optimising the existing model produces incremental efficiency. It does not produce structural fitness. And structural fitness — an operating model designed for the conditions that actually exist, not the conditions that existed in 1997 — is what the next few years will sort for.

This Note is part of the Articul8 AI and HR Operating Model series. The full structural reconception argument is developed across four Research Briefs — Understand the Disruption, Assess Your Position, Design the Response, and Govern the System — available in Briefs

An Articul8 Research Publication  ·  Chris Long, Founder Elev8 Group  ·  March 2026