Skills Transparency: The Key to Embracing AI in the Workplace

Source article: Why Employees Resist AI, and How Companies Can Win Them Over (HBS Working Knowledge, June 2026, based on Narayandas & Zhang working paper No. 26-050)

You are now royalty without a kingdom.

That is how Harvard professor Das Narayandas describes what happens to professionals when AI quietly hollows out their role. In a new working paper, Shunyuan Zhang and he, argue that at least 30% of generative AI projects will be abandoned, not because the technology falls short, but because the very people it was built for quietly reject it.

Their term for this is symbolic adoption: employees comply on the surface while undermining the tool underneath, because it threatens their identity, their status and their sense of self-worth. The paper names three mechanisms: role compression, control shift and span erosion. The diagnosis is sharp, and the human-centred remedies the authors propose deserve a wide audience. I agree with them.

And, I believe there is an essential piece missing from the conversation.

If professional identity is rooted in performing valuable work and carrying meaningful responsibility, then organisations should first make that work completely transparent.

Rather than describing jobs at a high level, deconstruct every role into its underlying skills and sub-skills: tasks, traits, tools and knowledge. A detailed skills architecture lets people see exactly what AI augments, what AI automates and, more importantly, which human capabilities remain indispensable.

That changes the narrative. Instead of experiencing "AI is replacing my job", employees see AI taking over specific tasks while significant room remains for human judgement: contextual reasoning, relationship building, creativity, ethical oversight, and the design, governance and continuous improvement of AI guardrails.  The perceived loss of control becomes a redistribution of control.

Humans remain responsible for defining objectives, setting boundaries, validating outcomes, managing exceptions and improving the way AI operates. AI becomes a capability inside a human-designed system, not a replacement for human value.

A comprehensive skills profile delivers another benefit: clarity. It lets organisations redesign roles with precision, redefine responsibilities and build transparent, skills-based career pathways. Narayandas and Zhang call for credible "redeployment pathways"; a skills architecture is what makes those pathways concrete. Redeployment becomes far less threatening when employees can see the skills they already possess, the adjacent skills they can develop, and the new places where they continue to create value.

This is why I’m convinced that the skills-based organisation is not simply another HR fad. It is becoming the organisational operating model for the AI era. Without a robust skills architecture, organisations struggle to understand what work can be augmented, what should remain human, where new capabilities are needed and how talent can move across the enterprise. With it, AI adoption becomes more transparent, more agile and, let's be honest, more human.

The real antidote to AI anxiety is skills transparency...

When people understand how their contribution evolves, rather than fearing it disappears, they embrace AI as an amplifier of their expertise instead of a threat to their professional identity.

Technology may change the work. Skills provide the language that helps people understand where they continue to belong, how they continue to contribute, and why they continue to matter.

That may well be the missing ingredient for successful AI adoption.