Most organisations now have AI. Very few are getting real value from it.

That’s the uncomfortable message coming out of the Gartner HR Symposium/Xpo — and it has big implications for HR leaders.

The pressure everyone feels (but few say out loud)

Boards and CEOs are asking a simple question:

“How do we grow revenue without growing headcount?”

At the same time:

  • AI investment is at record highs
  • Almost every company has rolled out tools
  • Productivity and ROI are… underwhelming

Gartner’s data is blunt:

  • Only 1 in 3 AI initiatives improves productivity
  • Only 1 in 5 delivers measurable ROI

So what’s going wrong?

The real problem: we added AI to broken work

Most companies focused on preparing people:

  • Training
  • Change management
  • Culture

What they didn’t do enough of: redesign the work itself.

If the underlying processes are slow, fragmented, or outdated, AI just helps people do the wrong work faster.

This is where HR’s role changes.

HR’s shift: from workforce manager to work designer

The new mandate isn’t just:

  • “Do people know how to use AI?”

It’s:

  • “What work actually needs to be done?”
  • “How should that work change in an AI-enabled organisation?”
  • “Where should humans still add value — and where shouldn’t they?”

That’s a strategic, business-level question. Not an HR side project.

Three ways work is really changing

Gartner breaks it down simply:

  1. Augmenting work (most common) AI makes existing work faster or easier Think copilots, drafting, analysis, summaries Low risk — but usually low return if left unfocused

  2. Re-engineering work (where value starts showing) Processes are redesigned end-to-end AI takes over large chunks of operational work Often frees capacity to reinvest in growth

  3. Inventing work (game-changing, but risky) Entirely new AI-native ways of operating Examples: fully automated operations, “dark factories” High risk, high payoff

The mistake is treating AI as one thing. Smart organisations manage all three — deliberately.

Why “letting people experiment” doesn’t work

Employees are already using AI. That’s not the issue.

The issue is misdirected usage:

  • Helpful, but not valuable
  • Convenient, but not strategic

The fix isn’t more training. It’s direction.

HR needs to guide teams to apply AI to:

  • Revenue bottlenecks
  • Customer friction
  • Cost drivers
  • Speed and quality constraints

When AI is pointed at real business pain, value shows up fast.

The HR + IT partnership is now mission-critical

This is no longer optional.

  • IT brings platforms, data, security
  • HR brings workflows, incentives, skills

If those two aren’t tightly aligned, AI investments stall — or worse, create risk.

Think of them as co-owners of how work gets done.

One surprising blocker: leadership mindset

A big finding from Gartner:

  • Most leaders think they’re supportive of AI
  • Their behaviour says otherwise

The most effective fix wasn’t training — it was visibility.

When leaders saw simple data showing how their AI adoption compared to peers, many asked for help on their own.

Sometimes the fastest way to change behaviour is just to hold up a mirror.

The takeaway for business leaders

AI won’t magically deliver productivity gains.

The winners will be organisations that:

  • Redesign work, not just deploy tools
  • Treat AI as an operating model shift
  • Position HR as an architect of work, not just a steward of talent

This is the quiet shift happening in 2026.

HR isn’t just supporting the business anymore. HR is shaping how the business actually runs.

Source: Gartner HR Symposium/Xpo keynote