How to become an AI savvy leader
Five moves that balance AI opportunity with safety.
In brief
- Leading on AI means deciding in a field you are not expert in, heading somewhere none of us can clearly see yet.
- Five moves, from the practical (secure tools, a small team, a readable policy) to the personal (educating yourself, leading by example).
- The job is to hold opportunity and safety at once, visibly.
By Inbal Rodnay
Leading without a map
It is hard to lead on AI.
We like leading from expertise. We know our profession, we know our clients and our internal stakeholders, we have done the work ourselves and we can tell good from bad at a glance.
With AI, the path keeps shifting. The tools change, the expectations change, the risks change. Nobody knows the destination, and we are still expected to lead our teams there safely.
Your people are probably further ahead than you think. In McKinsey’s 2026 global survey, 70% of employees and leaders surveyed said they felt personally ready to use AI, but only 27% of leaders said their organisations were ready to make the changes needed. In Australia, 44% of employees say they have used AI in ways that contravene policies and guidelines (KPMG and University of Melbourne, 2025). In New Zealand, the figure is 34%. The question is not whether your team is using AI. It is whether your organisation is keeping up with them.
The point is not to become certain. The point is to lead in a way that is practical, visible and well managed, so you can keep moving.
1. Give your team secure AI tools
Start here. Everything else sits on top.
Your staff want to get the work done, and they need the tools to do it. If the business has not given them a secure workplace AI tool, many will use whatever they can access. That is shadow AI. It means you may not know what tools they are using, what data they are entering, or how that data is handled.
A workplace AI tool may involve an additional cost. But the right tool, with appropriate data protections and settings, can help you operate with confidence. Your organisation can manage access, and your team builds skills in technology that is fast becoming core to their work.
“AI can’t do this” complaints often turn out to be IT settings, not the tool. A feature may be switched off, or a connection to another system may not be enabled. Before you conclude the tool is limited, check what your own environment allows.
Choosing the tool is a decision of its own.
See: Five questions to choose your primary AI tool. For the data side, see: Is my data safe to use with this AI tool?
2. Nominate a small team to bring the knowledge home
You do not have to become the AI expert. But you do have to make sure that your team has access to the knowledge it needs.
Pick one person or a small group, and give them a clear mandate to learn this properly and bring the knowledge back to everyone else. In many teams, they are called the AI lab, the AI task force or the AI committee.
They try the tools, they watch what is changing, and they tell you what is ready and what is hype.
Make it a small team, not a single champion. As CA ANZ and ACCA’s 2026 research puts it: “If one self-taught individual leaves – the capability leaves with them.”
Include people who do the hands-on work, as they are often the first to see where AI gets things wrong, and someone who understands risk, data and IT. Then make one senior person accountable for how AI is used across the organisation. Accountability is the first practice in the Australian Government’s Guidance for AI Adoption, and New Zealand’s Responsible AI Guidance for Businesses covers it too.
This commitment often makes the difference. A small, committed team can get further than a much larger organisation where no one has been given the job.
3. Read your AI policy and make sure you understand it
Not skim it. Understand it. If you cannot explain a section of it, your staff probably cannot either, and that section needs rewriting, or raising with whoever owns the policy. A policy no one understands is a document that ticks a box, not a real guardrail.
A good policy has to cover more than the data basics. It needs to speak to agentic AI, the tools that can take actions on your behalf, and to AI that can click around your browser to fill out forms.
If your current policy stops at "do not put confidential information into public tools", it is likely out of date.
No policy yet? Then that is your first step. You do not have to write it from scratch or become a policy expert. See: Build your AI policy: a step-by-step method.
4. Educate yourself
Your own fluency matters more than you might expect. In McKinsey’s 2026 global research, organisations whose leadership teams had high AI fluency were nearly four times as likely to report real value from AI (35% versus 9%).
When you understand the technology, you can ask the right questions when your team wants to build an AI agent, or when a vendor offers to sell you an AI tool. You do not need to go deep into the technology. You need enough fluency to understand the capabilities, the risks, and what is coming.
It also helps you judge what you actually need. Some vendors now offer “AI employee” products. Before you buy, ask three questions. Could your team do this inside a platform you already pay for? Is this likely to become a standard feature within months? And where will your data go? See: Is my data safe to use with this AI tool?
CA ANZ’s AI Fluency Playbook and CA ANZ capability+ model are built for this, and they give you and your team a shared starting point. They pair well with knowing where your team or organisation sits today. See: Where are you on the AI adoption curve?
5. Lead by example
This is where culture becomes guardrails. It is also free, and you can start today.
If you are not a heavy AI user, make a point of trying one task a day. Ask yourself, "How could I do this with AI?". If you are not sure, ask your organisation’s approved AI tool how it could help.
Then tell your team. Share what you tried. If it worked well, they will learn from it. If it was basic, they will feel validated and experiment more openly themselves.
The opposite is also true. If you treat your own AI use as something to hide, so will everyone else. Silence is the default. In Karbon’s State of AI in Accounting 2026 report, a global survey of accounting professionals, people underestimated how often their peers use AI by 29%. Only 18% thought their colleagues were excited about it, far below the real figure. When leaders talk openly, they close that gap. There is no shame in using AI appropriately. Encourage people to share what they are learning, including mistakes, and to check outputs before using them. Model the smart way.
Managers are the lever here. Train your managers to encourage wise use of AI. When work lands on their desk, they can ask, where appropriate: “Did you use AI to help check this, and how did you assess its suggestions?” Making this a habit can help improve the quality of work.
See: How to check AI's work.
Culture is a governance input. The way you behave with AI becomes the norm, and it can shape behaviour as strongly as any policy.
Holding both, on purpose
These five moves don’t resolve the tension between opportunity and safety. The pull between "do not get left behind" and "keep us safe" is real.
A leader who only chases opportunity leaves people anxious and the organisation exposed.
A leader who only manages risk is seen as a blocker.
A savvy leader can hold both, openly.
Three decisions only you can make
1. What happens to the time AI saves? Time saved does not automatically become value. Xero’s 2026 research found AI already saves Australian practices about a working day a week, yet only 44% are directing that time into higher-value advisory work. Decide where the hours go, whether that’s better analysis, advice, development or growth, and tell your team. The same question applies in-house, whether the hours go to analysis, business partnering or faster reporting. When people know saved time won’t simply be filled with more of the same work, they are more open about how they use AI.
2. How will your people still build judgement? Accountants have traditionally built judgement through the groundwork: reconciliations, first drafts, variance analysis. As AI takes on more of it, that learning has to come from somewhere else. This is referred to as the ‘experience paradox’. Decide how your people will still learn to spot what looks wrong, for example by reviewing AI output against source records and explaining why they agree or disagree.
3. Who is accountable when AI gets it wrong? Accountability cannot be delegated to technology. In CA ANZ and ACCA’s 2026 research, 71% of finance professionals agreed they should be held responsible for the AI outputs they use. Yet 67% were highly concerned about relying on AI insights that can’t be independently verified, and only 9% felt well prepared to manage that risk. Make it explicit who reviews and signs off AI-assisted work, and what checking is expected.
Where to start
If you find this hard, you are not alone.
And you are still the right person to lead it.
Pick one of the five moves and take one practical step on it this week. If you are unsure where to begin, start with the first: secure tools.
Sources
McKinsey & Company, From adoption to impact: three horizons of AI transformation, July 2026
Karbon, The State of AI in Accounting 2026
Xero, Modern Practice Playbook, September 2026
Australian Government, National AI Centre, Guidance for AI Adoption, 2025
Disclaimer
This guide provides general guidance only and is current as at October 2026. Tool names, features and capabilities were correct at that date but may change. The publication therefore focuses on general principles rather than specific feature lists.
This guide is not intended to be comprehensive and is not a substitute for professional, legal or other advice tailored to your circumstances. Use it alongside your organisation’s artificial intelligence policies and your applicable legal, regulatory, professional and ethical obligations. Check current vendor terms, product information and your own obligations before acting on or relying on the information.
While Chartered Accountants Australia and New Zealand (CA ANZ) seeks to ensure that the information is accurate and current, it does not guarantee its accuracy, completeness, currency or suitability for any particular purpose. To the fullest extent permitted by law, CA ANZ disclaims all warranties and liability for any direct, indirect, special or consequential loss or damage arising out of or in connection with any inaccurate statement, opinion, advice or information contained in this publication, any omission from it, or the use of or reliance on this [guide/document] or any material it contains.
© 2026 Chartered Accountants Australia and New Zealand ABN 50 084 642 571