From AI adoption to transformation: Why HR is redesigning work through a modern stack

Most organisations today are experimenting with AI, but far fewer are using it to rethink how work gets done fundamentally. As per research, while 88% of organisations report using AI in at least one business function, only about one-third have begun scaling AI across the enterprise. 

That distinction sat at the heart of a conversation at People Matters Tech HR Singapore 2026, where Mukul Jain of HONO and Jolene Huang of Publicis Groupe challenged a common assumption in enterprise AI: that implementing tools and automating workflows naturally leads to transformation.

The discussion argued otherwise. AI adoption may improve efficiency, but transformation requires something deeper. It requires rethinking HR’s operating system, moving from isolated use cases to a more modern, intelligent stack, one where workflows, data, skills and decision-making work as an integrated ecosystem.

In that framing, AI stops being an additional layer on top of legacy processes and becomes part of how work is redesigned.

Automation improves tasks. Transformation changes outcomes

A core theme in the session was the distinction between automation and transformation.

As Mukul Jain framed it, using AI in fragments as a transactional layer is fundamentally different from using it to generate insights, support decisions and improve business outcomes. That is where transformation begins.

For many organisations, AI adoption has first delivered productivity gains. Routine queries are automated, repetitive effort is reduced, and execution is accelerated. But efficiency alone does not amount to transformation.

The difference lies in whether AI is simply improving existing tasks or changing how decisions and workflows operate.

Examples shared by Jolene Huang illustrated that distinction. At Publicis Groupe, AI agents are being used not only to automate support tasks but to augment managerial judgement, from helping leaders prepare for difficult conversations to navigating employment law questions across Southeast Asia. These applications point to AI supporting judgment, not simply replacing effort.

The same principle was visible in Publicis’s AI assistant for employee queries. While reducing repetitive administrative work, the broader value lay in releasing HR capacity for more strategic contributions.

The implication is that automation may remove friction, but transformation happens when AI changes the structure and quality of work.

Transformation starts with redesigning workflows.

One of the strongest insights from the conversation was that transformation did not begin with deploying more technology. It began with examining work.

Jolene described a journey at Publicis Groupe that began with experimentation.

This included internal challenges, encouraging teams to automate parts of their work and build familiarity with AI. But the bigger shift came when attention moved from experimenting with tools to mapping workflows and redesigning processes.

That meant asking more structural questions, such as where work can be automated and where AI can augment decision-making. And where does human intervention remain essential? 

This sequencing matters. Many organisations still approach AI tool-first, looking for use cases after deployment. 

What emerged in the session was almost the reverse logic: transformation starts by understanding how work should operate differently, then applying technology to support that model. 

For HR, this shifts AI from a technology initiative to an organisation design question. Here, the idea of a modern HR stack becomes more than infrastructure; it becomes an enabler of transformation. 

The bigger barrier is capability, not technology

A recurring theme in the discussion was that resistance to AI often stems less from technological limitations and more from uncertainty, particularly fears around displacement.

But the session pushed back on a simplistic jobs-replacement narrative. Instead, the discussion focused on how roles are evolving and being repurposed as new forms of expertise emerge, from automation fluency to agent-building and systems thinking.

Seen this way, the challenge is not whether work changes, but whether organisations are helping people evolve with that change.

That was reflected in examples such as Publicis’ virtual AI playground, embedded AI coaches and continuous experimentation models, positioned not as isolated innovation efforts but as infrastructure for building adoption at scale.

The conversation also addressed a growing concern around overreliance on AI and the risk of diminished human thinking. Here too, the emphasis was not on choosing between human judgment and AI, but on using technology to expand capability while continuing to upskill and evolve.

HR’s opportunity is bigger than adoption.

Taken together, the session pointed to a broader shift underway.

The next phase of AI transformation may be defined less by deploying more tools and more by how organisations redesign workflows, combine human judgement with machine capability, and build the skills needed to sustain that shift.

That creates an opportunity for HR to not only support AI adoption but also to shape the systems, skills and work models through which adoption creates value.

And that may be the bigger implication of this shift. As organisations move from experimenting with AI to embedding it into how work operates, competitive advantage may depend less on access to technology itself and more on how effectively work is redesigned around it.

The move from adoption to transformation, as the session suggested, is ultimately not about using AI to accelerate old ways of working, but about rethinking work itself.