Start With Your Data, the Rest Will Follow

"What do you mean we can't use AI in our business?"

This is a common — and often incredulous — reaction I hear from prospective clients when I tell them that they can't use AI in their business. The surprise typically stems from a belief that they can simply point an AI at their data and it will magically figure out what to do and solve any problem.

Unfortunately, AI isn't magic, it’s maths.

If a human can't make sense of your data, it’s very unlikely that a machine can. AI is powerful, but it isn't omniscient. It relies entirely on the quality, structure, and context of the data it's given. And if your current reports are riddled with gaps, inconsistencies, or stored in ten different systems that don’t talk to each other — AI isn't going to unify that chaos for you.

Start with the Foundation: A Data Strategy

To have an AI strategy, you have to start with having a data strategy. That means asking some unglamorous, but critical questions:

• Where is your data stored?

• What format is it stored in?

• Is it machine readable?

• Has your data been tagged to give it some form of structure?

• Do you have a cloud-based infrastructure that supports high availability and scalable data operations?

If you can’t answer those questions confidently, you’re not ready for AI. But that doesn’t mean all hope is lost — far from it. Each of these issues is solvable. There are mature technologies, frameworks, and best practices to help businesses get their data into shape. It just requires a pragmatic approach and a commitment to investing in your data the way you'd invest in any critical business infrastructure.

AI Shouldn’t Be a Solution in Search of a Problem

Another pitfall I often see is businesses wanting to “use AI” without a clear intended outcome. There’s a vague sense that it’ll make things better, faster, or cheaper — and in the right context, it absolutely can — but dropping AI into your business with no focus is like hiring a team of engineers before you know what you're building.

Start by asking:

• Where does your business hurt the most?

• Which areas are constant bottlenecks?

• Are there processes that are manual and repetitive?

• Where are people spending time doing low-value tasks?

• Which tasks require your best people ‘to go dark’ once a month in order to get them over the line?

These pain points can guide you toward specific opportunities where automation, machine learning, or generative AI could deliver real ROI. But you need to tie your AI aspirations to your operational and data reality.

The Magic Is In The Data

Another myth is that AI can make sense of any data, no matter how messy. And while AI is surprisingly resilient — it’s not a miracle worker. There's a big difference between data that's a bit noisy and data that's outright chaotic.

I’ve seen examples where the same data appears in five systems under slightly different names, and where multiple Excel files are being used to aggregate the total risk of a £2 Billion fund, or my personal favourite being that one user who was running the entire trading system on their personal laptop and who had not been on holiday for 2 years!

The real magic isn’t in the AI — it’s in the data. Clean, well-structured, accessible data is what makes intelligent systems work. Before you dive into models or automation, make sure your business can reliably capture, organise, and access the information that matters most.

AI Is a Team Sport

Even when the data is available and in decent shape, many businesses don’t have the internal culture or cross-functional collaboration needed to make AI succeed.

You need a technical team (data engineers, developers, analysts) working closely with the people on the front lines of the business — the portfolio managers, sales and trading and operations staff. They’re the ones who know where the problems live and what success actually looks like.

If AI initiatives are being developed in isolation by a technical team without real stakeholder input, you risk building something that no one uses — or worse, something that solves the wrong problems, but that still costs your business money!

AI Should Amplify People, Not Replace Them

There’s also a fear — or sometimes a hope — that AI will replace entire roles. But in practice, AI is far more valuable as an enabler rather than a replacement.

In corporate research, for example, AI can help quickly find relevant evidence, flag anomalies or generate reports — but a human still needs to validate and interpret those findings.

When implemented thoughtfully, AI can free your people from the drudgery of repetitive tasks and empower them to focus on higher-value work. That’s where the real productivity gains come from.

Final Thoughts: Be AI-Ready, Not AI-Hyped

AI is not a shortcut. It's a multiplier. If your processes are solid, your data is structured, and your teams are aligned — AI can supercharge what you’re already doing. But if you’re hoping it will clean up your data, fix your broken processes, or deliver clarity out of chaos… it won’t.

So before jumping headfirst into AI, ask yourself this:

Are you ready for AI, or just hoping AI will get you ready?

Start with your data, the rest will follow.