By Dominic Gopal
Ask most senior leaders how they use AI and you’ll hear a list of tasks: drafting, summarising, coding, formatting. Useful work — but all of it the same kind of work. They are using AI as an assistant. Almost none are using it as a thought partner. That gap is the difference between getting 10% out of these tools and the 90% that actually moves a decision.
The distinction is simple. An assistant executes; a thought partner challenges. An assistant writes what you tell it. A thought partner asks why you believe what you believe. One makes you faster. The other makes you right.
That framing comes from Geoff Woods’ The AI-Driven Leader. It only meant something to me when I tested it on a difficult client situation — the kind I would normally handle on instinct and a little emotion. Instead of asking AI to draft my reply, I spent an hour asking it to interrogate me. What is the other party actually afraid of? What am I optimising for — short-term comfort, or long-term trust? Which of my assumptions don’t survive scrutiny?
What I eventually sent was nothing like what I would have written alone. Calmer. Sharper. Aligned to the outcome I actually wanted, rather than the one my mood was steering me toward. The tool hadn’t handed me an answer. It had handed me better questions.
This is where our profession should pay attention, because challenge is already our trade. “Which of my assumptions don’t survive scrutiny?” is not a clever prompt — it is the auditor’s instinct, turned inward. We are trained to test everyone’s reasoning except, too often, our own in the moment we commit to a number, an opinion, or a position. A thought partner closes that gap.
And the gap compounds. A leader’s decisions accumulate faster than the people around them can correct course. A thought partner inserts one round of questioning between instinct and action — long enough for a weak assumption to be named before it hardens into a commitment. That single pause is the most underused source of judgement available to us, and it now costs almost nothing.
So treat AI less like a clerk and more like the sharpest sceptic in the room. The next time you reach for it to produce something — a memo, a recommendation, a reply you’ve half-written in your head — stop. Don’t ask for the answer. Ask it to find the holes in yours. Ask what you’re optimising for, and what you’re avoiding.
If you’re still using AI only to execute, you’re leaving the most valuable 90% on the table — the part that doesn’t do your work, but improves your thinking.
What’s a decision you made recently that a thought partner could have made better, if you’d let it?
Dominic Gopal is Founder and Chief Executive Officer of Hypernix Sdn. Bhd.
The content in this article is the personal view of the author and does not purport to reflect the views of Hypernix Sdn. Bhd. or the Malaysian Institute of Accountants.
