Why AI adoption is a people change challenge

AI is already in the building with seventy-one per cent of professionals using it at least a few times a week, and 53% saying it’s been implemented in workflows or embedded into organisational strategy.

So, job done?

Not quite.

In Talent’s recent webinar, From use to impact: How is AI really changing the way we work?, JP Browne was joined by Emily Zhang, Founder and Director of The HumAIn Impact, and Jack Jorgensen, General Manager – Data, AI & Innovation at Avec.

Drawing on insights from 1,505 business leaders and technology professionals across Australia and New Zealand, they explored why access to AI is moving much faster than the people, processes and structures needed to make it work.

Here are five takeaways for People, HR, transformation and workforce leaders.

1. Access isn’t adoption

Giving everyone an AI login and pointing them towards a training module might increase usage, but it doesn’t guarantee at all that the technology is helping anyone do better work.

People may know which button to press without really knowing:

• When AI is appropriate for the task
• What information they can safely enter
• How to assess the output
• When human review is required
• What they remain accountable for

As Emily explained:

“Training is an input. Capability and adoption are the outcomes.”

2. Managers are where the strategy gets real

Managers are usually the ones left translating broad organisational policy into everyday decisions, and when only 25% of managers surveyed feel ‘very prepared’ to manage AI use within their teams… That’s a problem.

Can the team use this tool? Is this task appropriate to automate? Who checks the work? What happens when something goes wrong? Where should the time saved go?

Managers don’t need to become AI engineers, but they do need enough clarity and confidence to guide their teams, challenge poor use and escalate the right issues.

Sending everyone a policy and hoping for the best? Probably not a winning change strategy.

Give your managers practical decision rights, clear escalation paths and examples grounded in the work their teams actually do.

3. The people doing the work should help redesign it

Leaders can set an AI strategy, but they don’t see every duplicated task, awkward handover or spreadsheet-shaped headache sitting inside each department’s workflows. The people doing the work within these teams, however, do.

Frontline employees can show you where time disappears, which processes create frustration and where an AI tool might genuinely help, and involving them early makes it more likely that the final solution fits the reality of the job.

Start with questions like:

• Where does the work regularly slow down?
• Which tasks create the most rework?
• What information is being entered more than once?
• Where are decisions getting stuck?
• What would a better outcome look like?

Then decide whether AI is the right intervention. Buying the tool first and searching for somewhere to put it is doing things backwards.

4. Guardrails need to be usable

Seventy-nine per cent of respondents say their organisation’s AI rules are clear. And yet only 16% are ‘very confident’ people understand what data can and can’t be entered into AI tools.

Something is clearly getting lost between policy and practice.

While a lengthy policy may satisfy a governance requirement, it won’t necessarily help someone make the right call halfway through a busy Tuesday afternoon.

People need practical guidance that answers:

• Which tools are approved?
• What information is off limits?
• Where must a human check the output?
• Who owns the final decision?
• How should a concern be raised?

The aim is what Emily described as “freedom within the frame”: enough room for people to experiment and improve the work, with boundaries that are easy to understand and apply.

5. Decide what happens to all the time saved

Eighty-five per cent say AI has made them more efficient, while 53% report meaningful time savings, and while that all sounds promising, saved time doesn’t magically redeploy itself.

If a task takes 30 minutes less, what happens next? Does that capacity improve customer service, lift quality, reduce pressure or allow people to focus on more valuable work? Or does it simply disappear into an already packed day?

This is where workforce design enters the chat.

Organisations need to consider how tasks, responsibilities and roles will change (not jump straight from “AI saved time” to “we need fewer people”).

AI may remove parts of a job without removing the need for the role. It may also create new requirements around reviewing outputs, managing exceptions, improving workflows and maintaining accountability.

Make the work better, not just faster

Successful AI adoption shouldn’t be measured by how many employees use a tool but by whether the organisation has improved the way work gets done.

And that takes more than technology. It takes managers who can lead the change, employees who help shape it, training connected to real roles and a plan for where the value will go.

Because maximum AI adoption was never the goal. Better work should be.

Explore the full conversation on AI adoption, capability and changing ways of working on YouTube.

Ready to align your people, roles and workforce strategy with what’s next? Talk to Solve.