Many organisations are introducing AI from the bottom up. Employees discover ChatGPT or another AI tool, experiment with it and gradually begin incorporating it into their work. Some quickly become highly capable users while others avoid it altogether. Managers may be clueless how extensively their teams are using it — or what they are using it for, which creates a significant problem.
One of the biggest risks for organisations is not simply inappropriate AI use.
It is failing to recognise where AI could genuinely improve the way work is done.
Leaders are often well placed to see across teams, processes and responsibilities. They understand where bottlenecks occur, where employees are overloaded and where inefficient processes have become accepted simply because “that’s how we’ve always done it”.
Empowering leaders to lead AI adoption
Undoubtedly, successful AI adoption needs leadership. And that means AI capability development needs to start at the top. Rest assured, however, leaders themselves do not need to become AI experts. They do not need highly technical knowledge before they can lead AI adoption.
They do, however, need sufficient understanding of LLM to ask good questions, identify opportunities, recognise risks and make informed decisions to assess:
- whether a proposed AI use case is genuinely useful
- where AI could reduce workload
- what risks need to be considered
- whether employees are relying too heavily on AI-generated outputs
- how AI might change existing roles and responsibilities
- what support their team needs to use AI effectively.
Leadership capability is therefore less about mastering every new AI tool and more about developing the judgement needed to lead in a workplace where AI is increasingly part of how work gets done.
Leaders set the tone for AI adoption
Leaders play a critical role in AI uptake as most employees take cues from their superiors. Most tend to fall into distinct categories:
- If leaders are dismissive of AI, employees may be reluctant to experiment.
- If leaders enthusiastically encourage AI use without discussing its limitations, employees may use it without sufficient caution.
- If leaders say nothing at all, employees are left to work it out for themselves.
- Leaders establish a culture where employees feel comfortable exploring AI while also understanding that its outputs must be questioned, checked and used with professional judgement.
The message should not be: “Use AI for everything.” Nor should it be: “AI is too risky, so avoid it.”
The better question is: “Where can AI genuinely help us do our work better, and how can we use it responsibly?”
Leaders need to understand the opportunities
AI can be a catalyst for broader conversations about how work is designed.
But leaders need enough AI literacy to recognise those opportunities.
With a practical understanding of AI, leaders can begin asking different questions.
- Why does this process take so long?
- Why are three people manually handling information that could be summarised or structured more efficiently?
- Why are employees repeatedly drafting similar communications from scratch?
- Why are we producing a report that nobody appears to use?
- Could AI assist with part of this process?
- Or, perhaps more importantly: do we need this process at all?
Leaders also need to understand the risks
AI enthusiasm without critical thinking can create its own problems. It should be well understood by now that LLMs can:
- Generate inaccurate information.
- Produce convincing responses that are completely wrong.
- Reflect bias.
- Lead to employees inadvertently entering sensitive information
- Generated generic, inaccurate or unsuitable content for intended audiences.
To counter these threats, leaders should understand issues such as:
- privacy and confidentiality
- hallucinations and inaccuracies
- bias
- intellectual property
- human oversight
- accountability
- over-reliance on AI
- appropriate and inappropriate use cases.
Importantly, understanding these risks should not be about creating fear. Rather, it should help leaders make better decisions not to eliminate all risk but to understand where the risks exist and put appropriate safeguards in place.
Leaders need to model good AI practice
If organisations want employees to use AI effectively, leaders should model what good use looks like by demonstrating how AI can help:
- prepare for a difficult conversation
- summarise complex information
- critique a draft
- identify gaps in a proposal
- generate questions before making a decision
- explore different perspectives on a problem
- streamline a repetitive administrative task
It also means modelling the habit of checking AI-generated outputs rather than accepting them at face value. When leaders demonstrate thoughtful and transparent AI use, they help shift AI from something employees experiment with privately to something teams can discuss openly.
That creates opportunities to share good practice, identify problems and learn from one another.
Start with leaders, then take capability to the teams
Leadership development is only the first stage.
Once leaders understand the opportunities, limitations and responsibilities associated with AI, they are better equipped to support the next stages of capability development.
They can help their teams build foundational AI skills.
They can identify workflows worth examining.
They can contribute to practical team-specific AI guidelines.
And they can create a culture where AI use is continually reviewed and improved.
That is why the first stage of my AI Capability Development Pathway is simple: Develop leaders first.
Not because leaders need to know everything about AI.
But because they need to know enough to lead others through a workplace that is changing rapidly.
AI capability cannot sit with a handful of enthusiastic employees.
If organisations want to use AI strategically, responsibly and effectively, their leaders need to be part of the journey from the beginning.