AI, Trust and the Power of Knowing Why

If AI is going to help lead our businesses, we need to know what it's thinking.

Ever sat in a meeting and had someone deliver a grand conclusion… without walking you through their thinking?

Frustrating, isn’t it?

Now imagine that “someone” is an AI system helping you choose your next regional lead, predict sales volumes, or decide whether to close a struggling store. You’d want more than just the answer. You’d want to know the logic behind it. The “why”.

That’s exactly the problem a new research project from Anthropic is trying to solve. It’s called Project VEND, and although the name sounds like it belongs to a dystopian thriller, the idea is refreshingly practical: get AI to explain itself.

And for those of us working in leadership, retail strategy, or executive hiring, this matters more than it might first seem.

From black box to breadcrumb trail

Right now, most large AI systems are brilliant at producing answers but they’re rubbish at telling you how they got there. It’s like hiring a consultant who gives you a slide deck, but refuses to take questions.

Project VEND changes that.

The researchers trained small language models to label their own reasoning as they go. Instead of just predicting the next word (as most models do), these models create little internal “thoughts” called VENDs breadcrumbs like:

• “I believe the customer base is declining.”

• “Therefore, store closure is likely.”

• “This assumes that regional footfall remains flat.”

That’s not just tech magic, it's potentially transformational. Because if we can understand what an AI is thinking, we can start to judge whether it’s thinking well. That opens the door to real trust, real collaboration… and much better decisions.

So… are we there yet?

Not quite. And this is important.

The researchers didn’t just shout about the project’s success but they openly acknowledged its failures.

Sometimes, the models trained using VEND performed worse than standard models. Sometimes their internal “thought bubbles” were wrong, oversimplified, or missing entirely. And sometimes the model gave the right answer for the wrong reason.

Which brings us to a vital point for all of us in leadership: AI isn't ready to replace human

reasoning.

It’s not ready to replace judgment, nuance, or experience. And in most cases, it still needs adult supervision.

And that’s okay.

In fact, it should be reassuring to retail leaders navigating this noisy AI boom. You don’t need to become an AI expert overnight. You don’t need to adopt every shiny tool. But you do need to stay informed,  because the pace is fast, and the implications are real.

What this means for retail leadership (and the talent we look for)

The retail sector is becoming increasingly data-driven. We’re not just asking AI to forecast demand, we’re asking it to shape customer journeys, optimise store formats, even weigh in on hiring decisions.

But speed and scale are nothing without trust. And trust needs clarity.

That’s why I found Project VEND so interesting not because it’s a finished product, but because it reflects the kind of leadership traits we look for in humans:

• The ability to articulate reasoning clearly

• The courage to own assumptions and biases

• The willingness to be transparent when things don’t go to plan

Whether we’re building AI systems or leadership teams, those qualities matter.

And yes, the machines are getting smarter. But the most effective leaders still know when to question the system. To chalenge the assumptions. To go beyond the output and ask: how did we get here?

Final thought

Retail has always been about reading between the lines, whether it’s consumer behaviour, team dynamics, or the signals buried in the data. The same goes for AI.

The next era of leadership won’t be about handing over decisions to algorithms. It’ll be about collaborating with them, interrogating them, and making sure they’re aligned with our values, goals and customers.

Project VEND is one small step in that direction. A tiny crack in the black box.

And that’s something worth watching.

If you’re interested in how technology is shaping leadership teams in retail and how we find people who thrive in this shifting landscape, feel free to get in touch. I’m always up for a chat.