This product has already been purchased.
- AI is moving beyond administrative support in the boardroom, with directors now using it to simulate hostile stakeholders, rehearse crisis scenarios and pressure test strategic decisions before meetings.
- Rather than replacing judgement, it is being used as a deliberate thinking tool to surface blind spots, challenge assumptions and test how strategies might fail in practice.
- As adoption grows, directors are also applying AI to risk exploration, communication testing and scenario planning.
AI isn’t just taking minutes anymore. Australia’s top directors are using it to role-play hostile investors, rehearse crisis scandals and even channel Warren Buffett before board meetings, turning it into a quirky, provocative and surprisingly practical strategic adviser.
Picture a board director the night before a pivotal strategy meeting. Concerned about the company’s financials, they give AI a chilling prompt: “If this company collapses in 18 months, write the post-mortem”.
In seconds, the machine delivers a brutally honest autopsy – missed warning signs, unchecked assumptions, reputational blind spots. As the director absorbs the insights, they walk into the boardroom armed not just with a broader perspective, but sharper questions.
This kind of unconventional use of AI is quietly reshaping modern governance.
Used well, AI can help boards to prepare more rigorously, interrogate risk, improve governance, sharpen communication and test assumptions before key decisions are made. Discipline matters, however, and AI doesn’t replace reading the board pack.
We asked board advisers and executives – including Tim Trumper GAICD, Matt Mueller GAICD and Adele Martin – for some unexpected ways they use AI.
Simulate the boardroom before the meeting
AI can act as a “virtual boardroom”, allowing directors to rehearse strategic discussions from different stakeholder perspectives before formal deliberation begins.
By assigning personas, directors can surface blind spots, test assumptions and challenge their own thinking from unfamiliar angles, says Trumper, shareholder and adviser to Quantium, chair of Maurice Blackburn Lawyers and author of AI: Game On.
“For example, ask ‘If I was head of risk at a top-tier investment bank, what questions might I have asked’,” he suggests.
“Or, when reviewing board papers: ‘If Warren Buffet was reading this, what three questions would he definitely ask?’”
The value, says Trumper, lies in deliberately stepping outside one’s own frame of reference. It introduces perspectives that are more adversarial, more disciplined and often more revealing.
Explore a proposal by asking “what if”...
From here, directors are increasingly using AI to interrogate whether decisions should be made at all.
Boards, after all, must balance shareholder obligations, regulatory compliance, reputational risk and customer expectations, but that doesn’t mean proposals can’t be pushed further.
To do this, Mueller, founder and MD of AI advisory Voleno, who also sits on four boards, suggests asking AI: “Even if we could, should we?” and “What are the risks if we’re wickedly successful?”
Directors can also use AI to generate the strongest possible counterargument to a proposal, especially when the system has sufficient context about the business, says Martin, founder of financial advisory Scale Squad.
Extend risk thinking beyond the obvious
As the use cases mature, AI is being used not just for decisions in front of boards, but for risks that are not yet visible.
By synthesising regulatory developments, market signals and emerging trends, it can generate plausible risk scenarios that extend beyond conventional forecasting.
Directors can use it to explore “adjacently possible” events – non-obvious chains of disruption that could materially affect the organisation – enabling them to benchmark against industry norms, detect weak signals earlier and ask sharper questions.
A simple but powerful prompt: “What would worry a sceptical investor here?”
Critique your own performance
The same logic of rehearsal is now being applied to personal performance. Namely, Martin uses ChatGPT voice mode to practise conversations such as sales calls or pitches.
“I’ll have it play an easy, medium or difficult client and give me direct feedback,” she says. “Saying something out loud is very different to reading it on paper – most people only discover that in a meeting.”
Trumper adopted a similar approach while chair of NRMA, using Claude to evaluate his AGM address through the eyes of customers, shareholders and financial journalists before refining the message based on the feedback.
“Pre- and post-meetings, I’d use AI to help me reflect on my own tone and style, and to be more aware of other perspectives to help the board synthesise complexity,” he adds.
Improve clarity and communication
Beyond rehearsal, AI is also being used as a blunt test of clarity.
Trumper agrees one of the simplest but most revealing exercises is asking AI to rewrite an annual report as a one-page explanation for a 12-year-old. If the message breaks down, it often signals deeper issues in clarity, structure or intent.
This kind of translation exercise is increasingly being used across shareholder communications, strategy updates and governance reporting as a way of stress testing whether messages are actually landing.
Use it as a crisis manager
In crisis scenarios, directors are turning to AI to simulate pressure environments before they happen.
Trumper suggests feeding AI the crisis management plan and telling it the situation that has occurred.
“Then ask it: ‘Give me advice of the next best actions I need to take as a director and review this from the position of Australian corporate law, the regulator, the customer and a journalist tracking the story’”, he suggests. “It will give you around-the-grounds advice.”
But he cautions this is a supplement, not a substitute. Outputs must be verified, context checked and sensitive data handled carefully.
“It is, of course, key to cross-check the advice and ensure the data is kept 100 per cent private, not used in training data with AI model settings,” he says. “It is also crucial you trust and verify any information, including that from AI itself.”
Other quirky ways to use AI
Compare governance policies against best-practice frameworks
Cross-reference disclosures and relationships for conflicts
Simplify complex financial or technical material into plain language
Combine macro trends into plausible future scenarios
Test the assumptions behind your strategy
Ask: “What biases might I be bringing into this decision?”
Get up to speed quickly on unfamiliar industries or technologies.
Most used AI tools
More and more executives are building personal “AI stacks” rather than relying on a single system.
Trumper describes a selective approach, using Claude for “white-collar, deep-thought, partnership work”, while also drawing on ChatGPT for creative input and Gemini for additional perspective.
He treats all outputs with caution. Noting AI can make errors, he regularly asks systems to validate claims and provide source material.
Mueller takes a broader, more experimental approach, working across Claude, OpenAI, Copilot, Gemini, Perplexity and advanced platforms like OpenRouter.
“There are currently 2.8 million AI models available, so I’m able to pull from the best capability for the insights required.”
However, he draws a clear boundary – he avoids using AI to summarise papers. “That feels like handing over an element of a director’s judgement and AI can’t see what’s not on the page or what was said before it. I’ll use it to pressure test a view, but the view is always mine.”
Martin focuses on practical execution. She uses ChatGPT, Claude and Canva to rapidly build dashboards and simple tracking tools, bypassing complex spreadsheet work.
She also uploads internal procedures and SOPs for easier reference, and uses tools such as NotebookLM to convert text into audio, making organisational knowledge more accessible across teams.
Latest news
Already a member?
Login to view this content