The Human Factor in Data and AI

Welcome back. Over the last couple of weeks, we have covered a lot of ground together. We started by tackling the core challenges of getting your data house in order, moved on to the foundational infrastructure needed to support real ambition, and most recently we delved into the bedrock of data governance. We have talked tech, we have talked pipelines, we have talked strategy.

But let us be direct: none of it matters without the right people.

You can have the most elegant Lakehouse architecture, the most pristine data models, and a governance framework that would make a regulator weep with joy. But if you have not addressed the human element, the teams who use the tools, the talent you are trying to attract, and the cultural transformation required to make it all stick, you have built a high-tech engine that is going nowhere.

This brings us to what I consider the most critical element in any data transformation programme: the Human Factor. This is where the magic happens, but it is also where the majority of data and AI initiatives ultimately stumble.

Data science, at its core, is a world of advanced mathematics, statistics, and uncompromising logical thinking. These are not, traditionally, the strongest traits you find in the people who excel at building client relationships, crafting compelling marketing narratives, or leading sales teams. Yet, in today’s business landscape, we are telling everyone they need to change their DNA and become “data-driven.”

This creates a fascinating, and frankly, dangerous paradox. On one hand, there is a genuine, widespread enthusiasm to understand and deploy AI. Everyone wants to be part of the future. The downside: this can breed an “ends justify the means” mentality. In the rush to get quick, visible results, teams often cut corners, bypass governance, and build solutions on shaky ground. It is a sprint for immediate validation in a world where, ironically, nobody believes anything will last long anyway.

But as leaders, our job is not to chase fleeting wins. It is to build a lasting capability. And that begins and ends with people. Based on my experience in the trenches, the entire challenge boils down to three core areas: Teams, Talent, and Transformation.

The Teams Conundrum: Breaking Down the Tower of Babel

One of the most persistent and damaging challenges in any data project is the failure of cross-functional teams to truly collaborate. It is less of a gap and more of a chasm. We bring together data scientists, data engineers, business stakeholders, IT security, and marketing experts, and we expect them to seamlessly create value. More often than not, it looks like a modern-day Tower of Babel run by Jira tickets.

Each group speaks its own jargon, operates on its own incentives, and carries its own biases.

I once worked with a major retail company that was incredibly excited to launch a new AI-powered recommendation engine. The data science team spent a few sprints building a technically magnificent model. The accuracy was off the charts. When it was deployed, however, engagement tanked.

Why? The conversations with the people on the shop floor were protocol-driven, basically the data team interviewing for requirements rather than engaging in the nuance of the customer journey and the sales journey (two very different beasts). Do not get me wrong: the model was a work of art. It was solving a different problem. When we got the technical people to actually spend a couple of days on the field, accompanying the sales team, having lunch together, talking to clients, using the legacy order app which the sales field officers had to cope with, everything changed.

That is how I believe we fix this. It is not about forcing everyone to become a data scientist. It is about building bridges, and this is what has worked for me:

I always start by creating dedicated translation layers. In my teams, we have had immense success by formally establishing roles like Data Product Owners who operate as Analytics Translators. These people are bilingual. They can speak the language of business and the language of data. They sit in on sales meetings and then turn around and translate that context into clear requirements for the engineering team. They are the essential connective tissue.

Nobody understands the whole complexity of a major enterprise project, that is why truly cross-functional teams are paramount. I do not mean a weekly check-in meeting. I mean putting a data scientist, a product manager, a marketing expert, and an engineer into the same virtual or physical pod, with a shared objective and shared accountability. When they are all responsible for the same outcome, not just their siloed part of the process, the conversations change dramatically.

The Talent Question: It Is Not About Hunting Unicorns

The second great human challenge is the talent gap in data and tech. The data and AI landscape is evolving so fast that organisations are in a perpetual state of catch-up. We face a dual problem: a shortage of technical expertise and a critical lack of the soft skills needed to drive change.

The common reaction is to go “unicorn hunting”, trying to find that mythical individual who is a world-class statistician, an expert coder, a brilliant communicator, and a savvy business strategist all rolled into one. Let me be clear: these people barely exist. And if you find one, you probably cannot afford them.

I saw this firsthand at a large insurance firm that wanted to use AI to help with claims processing. The most experienced claims adjusters, people with many years of invaluable institutional knowledge, were the most resistant. They were not anti-technology. They were afraid that a black-box algorithm would miss the subtle human nuances they knew were critical to spotting fraud or identifying a vulnerable customer. They feared their hard-won expertise was about to be rendered obsolete.

The project only succeeded when we changed the entire premise. Instead of trying to reinvent them, we reframed the initiative to empower them. We turned them into AI Supervisors. Their job became to train, validate, and correct the AI models, and provide fundamental feedback on the new systems, getting financial prizes for successfully deployed model versions. We leveraged their deep expertise to make the machine smarter. This did two things: it dramatically improved the model’s real-world accuracy, and it turned our biggest sceptics into our most passionate champions.

What worked, in practice:

Stop hunting unicorns. Start building balanced teams. You do not need every person to be good at everything. You need a team where skills are complementary: a great coder paired with a great communicator; a deep-thinking researcher paired with a pragmatic project manager. The team, as a whole, becomes the unicorn.

Invest in transformation, not just training. One-off workshops on Python or vector AI are not enough. You need to build a genuine culture of continuous learning. This means giving people the time and resources to learn. It means creating dual career paths, so your most brilliant senior data scientist can continue to grow as an individual contributor without being forced into a management role they do not want.

Recognise and reward the value of data roles. If you treat your data team like a back-office cost centre, do not be surprised when your best talent leaves for a company that treats them like a strategic asset.

The Real Transformation: Changing the Corporate DNA

This brings us to the ultimate human challenge: cultural transformation. I have said it time and time again, and I will keep saying it until it sticks: technology follows culture. You can spend millions on the most advanced data platform on the planet, but if your company’s DNA is still wired for gut-feel decisions, you have accomplished nothing. The platform will become a ghost town, an expensive piece of shelf-ware.

Changing the DNA of a company is not something a third-party contractor, nor a senior manager, nor a director, nor even a CTO can do alone. This is something that has to be driven by the CEO of the whole organisation, because everyone will have to change their mentality, processes, and ways of working.

The most powerful tool for cultural change is communication, specifically the celebration of small, concrete wins. When you roll out a new self-service dashboard for the finance team and they use it to identify a significant cost-saving opportunity in the first week, you need to make that story legendary. Share it in the company all-hands, write about it in the newsletter, have the CFO talk about it.

This does more than just make the finance or commercial teams feel good. It shows everyone else, in tangible terms, that this new “data-driven” way of working is not just more work, it delivers real, recognisable value. It proves that change is not something to be feared, but something to be embraced.

Technology, in many ways, is the easy part. The real challenge, and the greatest opportunity, lies in empowering our people to harness it.