What has shifted in the 8 months since I spent last year presenting to CPOs and their cohorts across industries

Where We Left Off

Throughout 2025, I had the opportunity to present to CPOs and their leadership cohorts across a wide range of industries: from financial services and professional services to healthcare, retail, and technology. Across those conversations, a single question kept surfacing: how should CPOs position themselves as AI moved from pilot to practice? The frameworks I shared offered a diagnostic lens, mapping how AI was entering organisations (Top-Down, Bottom-Up, Siloed, Organic), and how CPOs needed to shift from HR operators to strategic architects.

The core argument was this: the pace of AI change would outrun traditional organisational change cycles, and CPOs who did not act would find themselves reacting rather than shaping.

Eight months on, I thought it was worth coming back to that conversation. Not to restate the frameworks, but to be honest about what has shifted, what has played out as expected, and where the picture is more nuanced than the 2025 discussion suggested.

“The CHRO and CPO are becoming the new Chief Transformation Officers — whether they’re ready or not.”  — Deloitte Human Capital Trends

What Has Emerged in 8 Months

The pace of change has been consistent with what the research suggested, though the shape of that change has varied by sector, organisation size, and leadership maturity. Agentic AI — systems capable of reasoning across multi-step tasks with limited human direction — has moved from research and early pilots into active enterprise use cases. Regulatory frameworks have continued to develop. And the practical experience of leading AI adoption has added useful texture to what was, in mid-2025, still largely a strategic discussion.

This update identifies five areas where observations and practice have shifted meaningfully, and considers what that means for how CPOs approach the agenda from here.

01	AI adoption patterns have diversified: most organisations are now navigating more than one model simultaneously
02	The CPO’s strategic role is gaining definition, though it requires active assertion in most organisations
03	AI literacy gaps have become more visible and are beginning to affect team dynamics and productivity
04	Agentic AI has brought workforce transition timelines forward, making planning more concrete and more pressing
05	Governance has moved from a forward agenda item to an active operational requirement

SHIFT 01 — Adoption Patterns: From Distinct Pathways to Overlapping Reality

The four deployment models presented in 2025 — Top-Down (board or executive-led strategic direction), Bottom-Up (technology entering through third-party and vendor relationships), Siloed (function-specific pilots), and Organic/Decentralised (employee-driven experimentation with freedom to innovate) — were offered as a diagnostic lens to help CPOs identify how AI was entering their organisation.

Eight months on, the most consistent observation is that the majority of larger organisations are no longer sitting clearly in one model. They are operating across several simultaneously, often without a shared view of that reality across the leadership team. The practical question has shifted from ‘which pathway describes us’ to ‘how do we manage the interactions between pathways we are already in’.

What Has Shifted

THEME	JUNE / AUG 2025	FEBRUARY 2026
Top-Down	Board or executive mandate providing clear direction; AI strategy tied to defined strategic or financial priorities	Still the clearest governance model where it exists. In practice, the gap between executive intent and operational implementation has widened. Middle layers often lack the clarity or capability to translate top-down direction into practice
Bottom-Up	AI entering through vendor relationships and third-party tools; IT and procurement as primary gateway	The volume of available tools has increased substantially. Many organisations are finding that procurement controls are not keeping pace with the rate at which new tools are being evaluated and trialled by teams
Siloed	Individual functions piloting AI solutions to solve local challenges; contained and manageable	Siloed pilots remain common and often productive. The emerging challenge is translating successful function-level pilots into broader organisational capability without losing momentum or creating new inconsistencies
Organic / Decentralised	Employees experimenting freely; innovation as cultural norm; bottom-up energy seen as an asset	More prevalent than many organisations anticipated. Works well where there is cultural and capability maturity to support it. Where that foundation is thin, the absence of guidance around data, quality, and appropriate use is becoming more visible

The practical implication for CPOs is that most organisations need a coordinating view across all four pathways — not to centralise or slow adoption, but to ensure that People considerations (capability, trust, ethics, transition support) are visible wherever AI is entering. Without that, each pathway develops its own norms in isolation.

The CPO opportunity: Establishing a cross-pathway view of how AI is entering the organisation — and ensuring that workforce implications are part of the conversation at each entry point, not addressed retrospectively.

SHIFT 02 — The CPO Architect Role: Real, But Contested

The 2025 argument was that CPOs needed to evolve across four dimensions: Strategic Architect, Integration Orchestrator, Regulatory Navigator, and Technology Partner. That framing has proven directionally sound. The practical experience of eight months, however, is that the role requires active positioning. It does not simply emerge by virtue of the CPO’s organisational mandate.

In a number of organisations, CIOs and CTOs have naturally extended their remit to include workforce capability as part of technology strategy. CFOs are increasingly framing workforce efficiency conversations through an AI lens. Where the CPO has not established a clear point of view, decisions affecting people strategy are being made in adjacent functions — not through deliberate exclusion, but through momentum.

The Emerging Tension

Where CPOs are gaining traction  • Workforce ethics and AI governance conversations • Skills architecture and reskilling investment cases • Employee trust and sentiment during transition • Regulatory navigation (Fair Work, data privacy)

Where the CPO voice is less established  • AI investment decisions (often CIO-led) • Workforce efficiency timelines (often CFO-framed) • Tool selection and vendor governance (IT-led) • Business unit AI strategy (varied People function involvement)

A consistent pattern in organisations where CPOs have established clear strategic influence: they brought a concrete workforce perspective to the C-suite table — rather than waiting for an invitation to comment on decisions already in motion.

SHIFT 03 — AI Literacy: The Divide Has Widened

The three-tier literacy framework — Awareness, Application, Deep Skill — remains the right architecture. What has become clearer over eight months is that the rate of progression across tiers is uneven, and that the gap between those who have actively developed AI capability and those who have not is becoming more visible in everyday team dynamics.

THEME	JUNE / AUG 2025	FEBRUARY 2026
Awareness tier	Most employees had broadly similar, relatively low AI awareness; baseline roughly consistent across the workforce	A subset of employees have self-directed significant development. The overall floor has risen, but the range of capability within teams has widened noticeably
Application tier	Formal programs planned; sandboxes discussed as a near-term initiative	Employees who have reached application proficiency are creating measurable productivity differences. This is generating both opportunity and some peer friction around perceived fairness
Deep Skill tier	Small, specialist group; largely concentrated in technology functions	Demand for deep skill holders has grown substantially. Retention of this cohort is becoming a more active consideration for organisations that have invested in developing them
Multi-generational	Recognised as a factor; tailored support noted as important	Generational differences in AI comfort and adoption pace are showing up more consistently. This is less a conflict dynamic than a design challenge for learning and communication approaches

What Is Actually Working

Organisations that launched structured AI literacy programs in H2 2025 are beginning to see measurable returns. The critical design insight is that programs that were co-designed with employees, rather than delivered to them, have significantly higher engagement and application rates. The ‘sandbox’ model — contained experimentation environments — has also proven its value: teams that experimented without risk showed twice the rate of voluntary adoption.

What has not worked: one-off training events, compliance-framed AI awareness modules, and programs that treated AI literacy as a technology issue rather than a career development opportunity.

SHIFT 04 — Agentic AI: Bringing Workforce Transition into Focus

When the frameworks were presented in 2025, the workforce transition discussion centred on three pathways: role elimination, role evolution, and role creation. That framing was accurate, and the arrival of agentic AI in enterprise environments over the past eight months has given it considerably more texture and urgency than it had in the abstract.

Role Elimination	Some functions that were discussed as medium-term automation candidates are seeing that horizon compress. Routine analysis, first-line case management, structured reporting, and document processing are among the most affected, though pace varies significantly by sector and organisational context.
Role Evolution	The ‘role stretch’ concept from 2025 is proving accurate. Employees working alongside agentic AI are being expected to manage broader scope and greater ambiguity. The support structures to help people navigate this are still underdeveloped in most organisations.
Role Creation	New roles are emerging, though more gradually than some projections suggested. AI workflow design, human-AI team coordination, and ethics and audit functions are areas of growing demand. These require deliberate investment in pipeline and development.

“AI will not replace people. But people who know how to use AI — and organisations that enable them — will replace those that don’t.”  This observation from the 2025 whitepaper is gaining practical relevance. The organisations beginning to demonstrate this in measurable terms are those that started structural workforce planning early.

Workforce transition planning works best as a living process rather than a periodic project — one that maps which roles are evolving or at risk of automation, at what likely pace, and what redeployment or reskilling pathways are available. Organisations that have this in place are better positioned to make proactive decisions; those that do not are more likely to find themselves responding to events rather than shaping them.

SHIFT 05 — Governance: Moving from Principle to Practice

The 2025 frameworks identified governance as a critical CPO responsibility, framed around the dual risk of moving too slowly (competitive disadvantage) and too quickly (data, ethics, and trust failures). Eight months on, governance has moved from a planning consideration to an active operational requirement in most organisations.

The EU AI Act has entered enforcement. Australian privacy law amendments and Fair Work Act considerations around AI-influenced employment decisions are adding complexity that requires cross-functional attention. Shadow AI — employees using unsanctioned tools — is more widespread than many organisations had assumed, and the People function is well-placed to lead the response.

Common Governance Gaps Observed Across Organisations

• AI tool use policies that exist on paper but have not been effectively communicated or embedded in practice
• No clear mechanism for employees to raise questions or concerns about AI decisions that affect their work
• AI vendor selection processes that do not routinely include a People function perspective
• Performance frameworks that have not been updated to reflect what productivity means in AI-augmented roles
• No documented position on AI-generated work product — ownership, disclosure expectations, or quality standards

The Recalibrated CPO Agenda — February 2026

The three-horizon implementation model from 2025 (Mobilise 0–3 months; Strategise 3–12 months; Influence 12+ months) remains a useful structure. What the experience of eight months adds is greater clarity about what matters most within each horizon, and where the practical leverage points are.

NOW (0–3 months)	• Conduct an honest audit of AI tool use across the organisation — sanctioned and unsanctioned • Establish a workforce transition register: which roles, at what pace, what redeployment pathways exist • Bring a concrete people strategy perspective to the next C-suite AI discussion — with a point of view, not just questions • Identify the two or three AI literacy interventions with the highest near-term impact and resource them
STRATEGISE (3–12 months)	• Co-design AI policy with employees, not just for them • Update performance frameworks to recognise and reward AI-enabled contribution • Build the reskilling investment case with board-level metrics: capability gap closure, productivity gains, retention • Establish a quarterly workforce futures review that keeps the transition register current and connected to business planning
INFLUENCE (12+ months)	• Position People strategy as a primary lens in how AI investment decisions are evaluated at board level • Build external partnerships with education providers to develop pipeline for emerging roles • Engage in sector-level governance conversations — CPOs who help shape evolving norms will be better positioned than those who respond to them