Why HR must turn everyday AI experimentation into enterprise-wide impact
Across organisations, employees are quietly rewriting how work gets done.
They’re building custom GPTs to speed up admin. Using AI design tools to create presentations in minutes. Feeding survey data into language models to find emerging themes faster than any spreadsheet ever could. And just like Steven Frost notes in his recent video, many are “begging for forgiveness rather than asking for permission” when legacy tools slow them down.
This grassroots innovation isn’t a threat — it’s a signal.
A signal that employees are hungry for smarter, simpler, more human ways of working.
A signal that AI is already here, whether organisations formally adopt it or not.
And a signal that scattered AI experimentation is a goldmine for HR and leaders — if they can capture it, test it, and scale it safely.
Yet the 2026 People Priorities Report shows a worrying mismatch between AI appetite and AI adoption. While interest in AI continues to grow, 69% of HR teams say they are still in the “early experimentation” phase, and only a small minority use AI meaningfully in their workflows. Meanwhile, lack of internal expertise (69%) remains the biggest barrier.
So the question becomes:
How do we take employee-led AI ingenuity and evolve it into organisation-wide best practice — without exposing the business to risk?
Here’s how HR and leaders can do exactly that.
When employees create their own GPTs or find time-saving AI shortcuts, they’re doing what innovative people have always done: making work better.
Instead of shutting these experiments down, organisations should:
This openness also addresses a core issue highlighted in the People Priorities report:
employees lose trust when they feel unheard or excluded from shaping change.
AI adoption improves drastically when employees see that leadership values their ideas — not just the outputs.
Steven’s video emphasises the need for clear AI values, guardrails, and tool-testing frameworks before scaling any use case.
These guardrails should cover:
But here’s the key: guardrails must be agile, not a 60-page policy no one reads.
The most effective frameworks:
This enables teams to innovate while protecting the business.
Once employees surface great ideas, the next step is testing them before scaling.
A simple pilot framework might include:
Accuracy Testing
✔️❌ Does the AI tool generate consistent, reliable, fact-checked outputs?
✔️❌ Does it introduce bias?
✔️❌ Does it reflect organisational language and nuance?
In employee listening, for example, WorkBuzz’s People Science AI already conducts these checks, turning free-text into sentiment, themes, and clear “why” insights. This eliminates manual analysis while ensuring quality.
Risk Assessment
✔️❌ What data is involved?
✔️❌ Does the tool store input?
✔️❌ Is it compliant?
✔️❌ Are there GDPR implications?
User Experience Evaluation
✔️❌ Does using the tool actually save time?
✔️❌ Does it reduce frustration?
✔️❌ Does it improve accuracy or decision-making?
Scalability Check
✔️❌ Can this tool be adopted by many without intensive training or cost?
AI pilots should be short — typically 2–6 weeks — and designed to gather evidence, not perfection.
Too often, AI initiatives fail not because the tech isn’t good enough, but because communication is inconsistent.
The People Priorities report shows that where communication is effective, confidence in leadership is 58 points higher, and employees feel more connected to organisational priorities.
When rolling out approved AI practices:
Scaling isn’t just a technical process; it’s a cultural one.
Steven makes a powerful point:
HR must be centre stage in AI adoption because HR touches every employee.
HR is uniquely positioned to:
This aligns with the People Priorities report, which urges HR to take ownership of the employee experience, acting as the connective tissue between leaders and the workforce.
When HR leads AI adoption:
One of the simplest ways to scale what works is to deploy trusted, already-validated tools that minimise risk and reduce workload.
Tools like WorkBuzz’s People Science AI help HR teams:
This matters because only 21% of organisations listen to employees more than quarterly, partly due to shrinking HR capacity.
AI can free HR teams from data-crunching so they can focus on the strategic, human work of improving culture.
Employee-driven AI innovation is happening, whether leaders see it or not.
The organisations that win will be the ones who:
AI doesn’t replace people — it elevates them.
And when organisations harness the creativity already happening at the front line, AI becomes less about algorithms and more about people: freeing them, supporting them, and amplifying their impact.