Perspectives

Coffee with Kai: Why I believe most businesses are approaching AI the wrong way

Data and AI specialist Kai Henderson gives us a peek behind the office doors…

Kai explains why businesses should start with data - not AI - to unlock real value.

Kai Henderson originally started as our digital marketer and has since moved through a junior risk consultant role to commercial lead & data analyst and he’s now Marex’s data and AI consultant. Alongside his work at Marex, he is studying for a graduate apprenticeship in data science – and developing his taste for caffeine.

There’s a lot of noise around AI right now. Every week, another business is announcing a new tool, a pilot, a chatbot or a “transformation journey”. And don’t get me wrong, some of it is useful. But in my experience, most businesses are still asking the wrong question.

They’re asking: “What AI should we buy?”

I think they should be asking: “What problem are we actually trying to solve?”

That might sound like a small difference, but it changes everything.

At Marex, I’ve seen first-hand that the businesses getting the most value out of data and AI are not the ones chasing the flashiest demo. They’re the ones that understand their operations, know where the pain points are, and are willing to get the foundations right first. That means data engineering, data quality, governance, and a clear understanding of what the business is trying to achieve.

Because AI is only ever as good as the data underneath it.

If your data is spread across spreadsheets, systems, inboxes and people’s heads, then AI is going to struggle.

If your teams are using different versions of the truth, then the output will be unreliable.

And if nobody has thought about governance, there’s a real risk that employees are already putting sensitive information into tools they shouldn’t be using.

That is why I’m so focused on starting with data, not AI.

My offer is built around helping businesses identify where they can get the biggest return from data and AI, without wasting time and money on unnecessary complexity.

We begin with an audit of the current state: systems, data flows, manual processes, governance risks and opportunities for improvement. From there, we identify the highest-value quick wins, build a proof of concept where appropriate, and create a clear roadmap for the next 12 months.

The aim is simple: help businesses move from disconnected information and guesswork to connected, usable intelligence.

What makes Marex different is that we are not approaching this from a purely technical angle.

We’ve got decades of experience in sector-specific operational reality, and that matters.

It means we understand the commercial pressures, the regulatory environment and the practical constraints businesses are working within.

So when we recommend an AI or data solution, it is not just technically possible — it is relevant, realistic and worth doing.

That combination is rare. A lot of consultants can talk about AI. A lot of businesses can explain their problems. But far fewer can connect the two properly.

That’s where I think Marex has a real USP.

We understand the world our clients work in, and I understand the data and AI landscape well enough to know what is worth pursuing, what will create value, and what is likely to become an expensive distraction.

For me, that’s the whole point. AI should not be about chasing trends. It should be about making your business better.

If you want to chat about your business's approach to using AI based on its data, get in touch with Kai.

Perspectives