Our Blog

Earlier this year I was working with an AI assistant on marketing copy for a webinar. The copy was already good – well-structured, hitting the right notes for the audience. I knew it was good. And yet I kept going.

The AI kept offering more. Alternative phrasings. Refined word choices. A different angle on the subheading. Each suggestion was plausible. Many were genuinely better. And with each one I went a little deeper in, weighing options, revisiting decisions I had already made well.

An hour later, the copy was marginally better. What I had not done was give any further thought to the headline – which carries roughly 75% of the impact of any piece of communication. Or the subheading, which carries most of the rest. I had spent an hour optimising the body copy while the thing that actually determined whether anyone would read it at all stayed exactly as it was.

I had been gold-plating – adding unnecessary shine to something already fit for purpose. The work was done. The AI had encouraged me to keep going anyway.

Gold-plating already has a meaning in project management: adding features or refinements beyond what was needed, consuming time and budget without adding value. AI has given it a new, much more personal home. It’s not knowing when to stop – and while you keep refining, the thinking that would have created something genuinely new never happens.

This is the opposite problem to cognitive surrender, a risk we’ve written about before – where people stop thinking critically and simply accept whatever AI gives them. Here, you are still thinking. You’re just thinking about the wrong thing, at the wrong time, for far too long. 

Why AI pulls you in

AI is programmed to keep you engaged. That shows up first as narrowing: every suggestion is a plausible variation on what you’ve already done – never a prompt to step back and ask whether you’re working on the right thing at all.

There’s a second mechanism that makes this worse: sycophancy. AI is self-affirming by design – trained to validate rather than challenge, to agree rather than push back. A tool that never disagrees with you can never be the thing that makes you stop and rethink. OpenAI rolled back a model update in April 2025 specifically because it had become too agreeable. Researchers at Stanford found the same pattern across eleven major AI models in 2026.

Put the two together and you get a loop with no natural exit. Nothing inside the conversation ever tells you to stop.

If you’re a perfectionist, this is built for you

For a perfectionist, AI is both a dream and a trap. It hands you infinite justification to keep refining. Every suggestion is a plausible reason not to stop.

Before AI, that instinct had a natural brake. Producing another version took real effort, and eventually the effort itself made you stop. AI removes that brake completely. The cost of one more iteration is close to zero – and for a perfectionist, a near-zero cost of improvement is almost impossible to resist.

Researchers have started calling this ‘digital perfectionism’ – the fear of not being good enough, specifically amplified by how we interact with digital tools. A 2026 study found a direct link between this kind of perfectionism and AI dependence. The research so far is mostly in student populations, so the organisational evidence is still building – but the mechanism is well established, and it matches what I see in myself and in the leaders I work with.

A test that takes ten minutes

Here’s a simple way to check whether gold-plating is happening – in your own work or your team’s.

Take any piece of AI-assisted work and ask three questions:

• What was the strategic question this work was meant to answer?
• Did that question get more or less of your attention because of AI?
• Could this have been good enough thirty minutes ago?

If you said less to the second question and yes to the third, you may be guilty of gold-plating right now.

If that’s you, you’re inside the loop the same way I was – and the same way almost everyone using AI well is, at least some of the time. The difference is simply whether you notice.

Why leaders should care most

Leaders are particularly exposed to this. High status, high time pressure, real ego investment in work looking good. Exactly the conditions that make gold-plating most tempting – and exactly the conditions where it’s hardest to realise, let alone admit, that you’re doing it.

And it doesn’t stay private. Our emotional and cognitive states are genuinely contagious – research consistently shows that a leader’s internal state shapes the psychological climate of everyone around them. A leader who disappears into AI-assisted polishing, while presenting the result as strategic work, isn’t just costing themselves time. They’re teaching their team that this is what good AI use looks like.

Multiply that across an organisation and you get something worse than wasted hours: a culture that refines endlessly and originates rarely. Busier. More polished. Not actually thinking further ahead than it was a year ago.

The cost nobody sees on the balance sheet

None of this shows up as a problem. That’s what makes it dangerous.

Output keeps rising. Deliverables look sharper than ever. Multiply gold-plating across an organisation and you don’t get a visible failure – you get a collective ceiling.

Everyone is elaborating. Nobody stepping back to ask if we’re working on the right thing. The result is a business that looks more polished every quarter and thinks less originally every quarter – and because the polish is so visible and the lost thinking is so invisible, nobody notices until a competitor who never stopped originating pulls ahead.

What to do about it

Decide what good enough looks like before you open the AI tool – and ask whether this is the right thing to be working on at all, the thing that genuinely adds value. That single discipline creates a stopping point no amount of sycophantic encouragement can move.

And if you recognise yourself in this – as I did – the fix isn’t to use AI less. It’s to ask, more often than feels natural, what else needed your attention while you were polishing this.

The work was already good. The question is what didn’t you get to while you were making it shinier?

Continue the conversation

Join the live webinar

The AI Productivity Paradox: Why AI Adoption Isn’t Translating into Organisational Impact

Join Tom Flatau on 7 July for an interactive discussion exploring the research behind AI’s productivity paradox, the emergence of “System 3” thinking, and what leaders need to do differently.

Register for the Webinar

Book a Conversation with Tom

Tom works with organisations worldwide on AI adoption, leadership, mindset and organisational culture, helping leaders translate neuroscience into lasting behavioural and business performance change.

Schedule Now

Tom Flatau

CEO, Teamworking International  |  International Keynote Speaker

Tom has spent 25 years applying neuroscience to leadership and business. He works with organisations navigating AI adoption, leadership development and organisational change. Clients include BBC, HSBC, Unilever, Siemens, Emirates and Kier Group.