Fun fact, in 2023 I created and launched a Pilates reformer and wellbeing studio, Soulhub, in Sydney's Northern Beaches (if you are ever in Freshwater, come check us out www.soulhub.co).
I had zero experience in commercial fit-outs. Zero experience in the fitness/ wellness industry and zero experience in launching a business like this from scratch ( 🤔 interesting strategy you might be thinking). Fair to say, we are always learning, growing and trying new things.
Recently, we decided we would embrace AI to improve our efficiency. We added a bot to our client messaging. The purpose was to streamline bookings, confirmations, and the admin that eats away at my team's day. It made sense on paper. I was really excited.
Six weeks in, I wasn't looking at productivity gains. I was looking at angry client messages.
These long-standing customers, were now getting annoyed because they were dealing with a bot that couldn't be customised the way our business actually runs.
It was irritating them, instead of assisting them.
Every clunky, off-brand message chipped away at something we'd spent years building. Our reputation for being easy to do business with. We were jeopardising the trust and connection that we'd spent years building.
I'd handed a vital customer touchpoint where that connection gets built, or lost, to a tool that wasn't ready for it.
So I made the decision to pause the AI tool. It's not a complete no; it's just not a yes now. Not until we can customise it properly and protect the connection that's always mattered most.
This little hiccup, and talking with many of my corporate and government clients, got me thinking about trust in an AI world.
The businesses actually getting a return
The tech itself had real limits, it couldn't be customised the way our business runs, and that's a genuine product gap. But to be honest, that wasn't the real issue. The issue was I rolled it out without stress-testing it against our customers first.
That wasn't a technology failure. That was a leadership failure. And it was mine, because I failed to use my own judgement before I hit go.
Turns out we're not alone. Most businesses trying to make AI work are stuck in exactly the same place.
McKinsey's most recent State of AI research found that close to 90 percent of organisations are now using AI in some form. Only around 6 percent are actually seeing a real return. The rest are stuck in what they call a pilot loop, endless trials, nothing that sticks.
Leaders in that research were twice as likely to blame their people's readiness for the problem than to look at their own role as the leader. It's easier to point at the tool, or the team, than to ask yourself, what did I actually do with the tech once I had it, and how did I engage the team, or not, in the process? That was the mistake I made myself, and it had real implications for trust with my customers and my team.
None of that lets the tool off the hook, it clearly mattered in our case. But Google worked something else out years before any of us were talking about AI. They studied, in what became known as Project Oxygen, what made their own managers effective, not the tools those managers used, the managers themselves. Technical mastery ranked dead last, eighth out of eight. Judgement, coaching, and genuine interest in people topped the list.
The tools have changed enormously since then. What makes someone trustworthy, with your team or your customers, hasn't. My own Soulhub story proves the point from the other side, I had access to good judgement, I just didn't use it in that moment.
I am not an AI expert, and I'm not going to pretend to be one. What I know, from close to three decades of working with leaders and teams, is the part that decides whether any of this technology actually earns trust once it's implemented. That's not a technical skill. It's a human one, and it's the one I initially got wrong myself.
Why this isn't really about the technology
In all my conversations with teams and leaders, about how they're feeling at work, how connected they feel, or don't, to their leader or their organisation, the answer rarely hinges on the technology.
There's definitely a lot of grumbling about the enormous amount of change going on inside businesses right now. But it's not the change itself that creates the disconnection.
It's almost always some version of: do I actually believe the people running this place have my back?
The latest Edelman Trust Barometer puts a number on what I've been hearing in those rooms for years. Eighty per cent of employees rarely or never trust their leadership to prioritise people over profit when the two come into conflict. Seventy one per cent expect their employer to downplay or avoid the conversation about what AI actually means for their job.
The problem isn't the technology. AI just amplified where the trust was already thin.
This is the part leaders get wrong most often. They think an AI strategy, a town hall, a slide deck, will fix it. It won't, not on its own, because the distrust isn't about the strategy. It's about every moment before it, the moments people weren't consulted or brought along, where they were told one thing and watched something else happen, where a leader avoided the conversation about what any of this actually meant for their team.
When the difficult conversation doesn't happen, people don't wait patiently for an update. Their brain fills the gap themselves, usually with the worst-case scenario.
The three places trust is actually being tested
AI has highlighted three places trust is being tested, and only one of them is about the tech.
The first is trust in the tool. Whether it does what it says it will. Not my lane, and I've been upfront about that from the start.
My lane is what happens once the tool is available. Helping people build the confidence and judgement to actually utilise the tech in front of them. The workplace of today asks for a digital mindset, and here's the good news, that's something that can be developed, not something you either have or you don't.
I won't tell you which AI tool to use, or how it works under the hood. What I will teach you is how to brief it properly once you've chosen one, because what you get back is only ever as good as what you put in. That's why I built the B.R.I.E.F framework.
The second place trust is tested is in yourself. Whether you trust your own expertise enough to catch it when AI hands you something confidently, but completely wrong, because it will. The only thing standing between that and your reputation is whether you've briefed it well in the first place, and whether you use your judgement to catch it and course correct.
After the Soulhub experience, I built myself a T.R.U.S.T filter, a process to pressure test my own judgement. I run every high-stakes piece of work through it now before it goes out under my name or anyone else's. It's become one of the first things I share with the leaders and teams I work with, because this is the checkpoint most businesses are skipping entirely, and it's teachable.
The third is trust between people. Your colleagues. Your clients. Your team. And here's what Soulhub taught me the hard way. That third layer doesn't break because of AI. It breaks the moment you skip your own judgement and let something through that hasn't been made properly, unmistakably yours.
If you're a leader trying to work out how to hold onto trust and connection while everything else moves, that's the conversation I'm having with teams right now. Please reach out I'd love to hear what it looks like inside your business.
Trust and human connection were never good luck. They are a skill, and right now they might be the only ones AI can't automate out of your business.