3 tensions HR and the business need to resolve around AI

AI adoption is moving faster than most organizations can redesign the way people work.

That creates a growing tension between business leaders looking for productivity and value, and HR and L&D teams tasked with helping people adapt.

To explore where this tension is showing up, we spoke with three experts working directly on AI, workforce transformation, learning, and enablement: Marlene De Koning, Director of Workforce Transformation at PwC; Anna Svitak, Learning Advisor Safety at Shell; and Milica Sapic, Senior Enablement Architect, GTM at Personio.

Their perspectives point to three connected tensions. The challenge is not simply how to train employees to use AI tools. It is deciding what should change in the first place and making sure the organization supports those changes.

‍

1. The business wants AI value. HR is asked to deliver training.

Business leaders are asking increasingly direct questions about AI: Where will productivity improve? What work can be automated? How quickly will we see value?

HR and L&D are often brought in with a different question: How do we train our people to use these tools?

This creates a fundamental disconnect between the business case for AI and how organizations approach people development.

‍

The problem, she argues, is that HR and L&D can be brought into the process too late after the business has already decided which tools to implement.

“They are tasked with delivering training for the organization, and on the execution part before it's actually defined what people should do differently.” said Marlene De Koning

‍

That distinction matters. Training people to use AI is only useful if the organization has also defined how AI should change the work itself.

That means asking:

• Why are we implementing these tools?
• What should employees do differently?
• Which processes need to change?
• What should managers reinforce?
• How should performance be measured?

Without those answers, training risks becoming disconnected from the business transformation.

As Marlene De Koning puts it:

“If that work is not being redesigned and managers do not reinforce new behaviors and set KPIs on those new behaviors, organizations can maybe have very well trained employees. But employees can still return to the old way of working.”

‍

And that is where AI investment can fail to translate into business value.

The question for HR and L&D:
Are you being asked to train people for a new way of working that the organization has actually designed or simply to train people on a new tool?

‍

2. AI changes faster than traditional training can keep up.

Even when the organization knows what it wants employees to do differently, another challenge emerges: the technology itself keeps changing.

Anna Svitak, who works on enterprise-wide learning and culture programs at Shell, sees this tension directly.

“AI is just developing so incredibly fast that we are kind of building some of these trainings, we start to roll them out and at that point AI is already at a different point again.”

‍

For L&D teams, that creates a difficult cycle.

A training program takes time to design, validate, launch, and scale. AI tools can change during that process.

The natural response is to focus training on specific tools and use cases. But that can make learning obsolete almost as quickly as it is delivered.

‍

Those underlying behaviors are more durable than any particular feature or interface.

Employees need to know how to evaluate AI output, adapt their approach as tools change, recognize where human judgment is needed, and apply AI appropriately to their work.

Otherwise, learning can become too prescriptive.

As Anna puts it, people can end up “just copy-pasting what they've learned in a training.” That is a very different capability from being able to use AI effectively in a changing work environment.

For L&D, the implication is significant. The goal cannot be to create a definitive AI training that employees complete once.

Instead, learning needs to help people build behaviors that allow them to keep adapting as the technology evolves.

The question for HR and L&D:
Are your AI programs teaching people today's technology or helping them develop the judgment and behaviors to work with whatever comes next?

‍

3. AI changes the operating model but HR is not always in the room.

The third tension moves beyond learning altogether.

AI can change how work is divided, which tasks are automated, where decisions are made, and what different roles are responsible for.

That makes AI adoption an operating model question, not just a technology or training question.

Milica Sapic, Senior Enablement Architect, GTM at Personio, argues that this is an area businesses can overlook.

“Businesses may be neglecting to see the operating models, to revise the operating models.”

The business case for AI asks: How do we get value from this?

But there is a second question: How do we design the organization so that it can actually capture that value?

‍

That second point is particularly relevant for HR and L&D.

If decisions about new workflows, roles, responsibilities, and ways of working are made without HR and L&D, these teams can end up entering the process once the most important decisions have already been made.

Their role then becomes implementation: communicate the change, build the training, and support adoption.

But HR and L&D bring expertise that can shape the transformation earlier from understanding changing roles and capabilities to anticipating the behaviors and organizational conditions needed for new ways of working.

Milica Sapic sees an opportunity for HR and L&D to have a more active role in these conversations:

“I believe we do have enough experience and knowledge to also shape how the organization will be set.”

‍

That raises a bigger question about the role of L&D in AI transformation.

If AI is changing the design of work itself, should L&D only be preparing people for that change  or helping shape it?

The question for HR and L&D:
Are you involved early enough to influence how AI changes the organization, or are you brought in once the operating model has already been decided?

What these tensions mean for HR and L&D

Across all three perspectives, the same pattern emerges. The biggest challenge with AI is not simply getting people to use the technology. It is connecting technology, work, behavior, and organizational design.

Marlene's point is that training cannot compensate for work that has not been redesigned. Anna's is that tool-specific training cannot keep pace with AI's development. And Milica's is that HR and L&D need a seat in the conversations where new operating models are being designed.

Together, these tensions suggest a shift in how organizations should think about AI capability building.

The question is no longer simply:

“How do we train our people to use AI?”

It is:

“What needs to change in the way our people work and what capabilities, behaviors, management practices, and organizational structures will make that change stick?”

For HR and L&D, that means getting involved earlier in AI transformation, building capabilities that can evolve with the technology, and connecting learning to the actual redesign of work.