AI Succeeds When Trusted Data And Human Judgment Create Smarter Businesses

I actually started my career in telecoms, however, I generated keen interest in communication tech, data and Architecture when I was in the third year of my engineering degree. It was in the very early stages of my career at Samsung where I was fascinated on how tech changes people's lives and delivers business value. 

HSBC was a defining chapter in my leadership journey where I came across real business transformations driven by data & analytics across more than 20 markets. HSBC taught me what scale really means, legacy technology, regulation, fragmented data, competing priorities and thousands of people affected by the decisions you make. It taught me that successful transformation is as much about leadership, trust and operating models as it is about architecture. 

At BCG, my perspective expanded from transformation to value. Leading Data, GenAI and Governance globally, while working with senior executives and clients, reinforced a belief I hold strongly today i.e. AI investment without a clear path to business value is simply expensive experimentation. 

At Exadel, that thinking has evolved further. My focus today is on AI-led value creation, how we move beyond pilots and productivity tools to fundamentally rethink processes, decisions, products and business models. 

So, my journey has really been an evolution from building technology, to transforming enterprises, to creating value through AI. And that continues to shape how I lead today. 

In my view data strategy start with with the business objectives i.e. WHAT is the organisation trying to achieve in terms of revenue, growth, better customer experience, risk reduction or perhaps an entirely new business model? Only then do I start to fit the data and AI and define the role it will play in getting us there. 

One lesson from leading large transformations is that it is very easy to create an impressive technology roadmap that delivers very little business value. I therefore work backwards from a small number of measurable business outcomes and connect them to the data, AI capabilities, architecture, governance and operating model required to deliver them.

Although it’s hard, I also believe strongly in measurable business value and tracking it from day one. Every initiative should have a clear line of sight to an outcome and someone in the business should define and own that outcome. Data teams should not be celebrating how many platforms they built or datasets they migrated, rather they should be showcasing who delivered the business value through the platform and lines of data transformation code they have written . 

Ultimately, a good data strategy should not feel like a separate technology strategy. It should simply be the business strategy, enabled by data and AI. 

I really like this question as the answer you mostly hear is that legacy technology is the biggest challenge. In my view It is the legacy thinking and complexity in the change management that have grown around it. 

I have seen organisations invest heavily in cloud, data platforms and now AI, yet carry forward the same fragmented processes, operating models and (no)ownership structures. Moving legacy data and the workload to the cloud and then adding AI on top of fragmented data does not make organizations modern and AI-ready. 

At HSBC, BCG and now Exadel, one lesson has been consistent i.e. modernisation works when you address technology, data, business/IT processes, operating model and people together. 

The other challenge is balancing transformation with business continuity. Large enterprises cannot stop operating while they modernise. The art is therefore to create value incrementally while progressively removing legacy complexity. 

My principle is simple, don’t modernise technology for the sake of being modern. Modernise what prevents the business from moving faster, making better decisions and creating value. 

Most large organisations already have enormous amounts of data. The problem is that it sits across different platforms, functions and business definitions. Simply bringing all of that data into one platform does not solve the problem. True value of the data is realized when it’s easily discoverable, and consumable and once consumed it should be connected, have clear business meaning/definitions and have enough context to understand how it relates. 

The real opportunity is to create a trusted and connected data foundation with common business semantics, clear ownership and governance so that people and increasingly AI agents can understand not only what the data says, but what it means in the context of the business. 

But I would take it one step further. Insight itself is not the end goal. An insight only creates value when it changes a decision or triggers an action. The most successful organisations will therefore connect data, analytics and AI directly into business workflows shortening the distance between data, insight, decision and action.

That, for me, is when data stops being an enterprise asset in theory and starts becoming one in practice. 

It’s important to first understand what data foundations are and for me it means a well defined enterprise data & AI platform with the right choice of tools/tech, north start Arch and clear roadmap for building quality and trusted data products. Strong data foundations are becoming even more important in the age of AI. You cannot build intelligent applications or increasingly autonomous AI agents on data that the organisation itself cannot trust, understand or access. 

But I believe we need to move beyond simply building data platforms. The real shift is towards data products connecting both structured and unstructured data with knowledge graphs, designed around specific decisions and outcomes. A customer data product, for example, should provide a consistent understanding of the customer that can be reused across sales, service, risk, analytics and AI rather than each function creating its own version of the truth. 

One of my biggest lessons is that transformation is not a technology programme. It is an organisational change programme, driven by business goals and enabled by technology. 

At HSBC, leading transformation across more than 20 markets taught me the importance of executive sponsorship, securing long term investment and strong foundations, while continuing to operate in a complex and highly regulated environment. At BCG, I saw another dimension that transformation only succeeds when there is absolute clarity on the business outcomes you are trying to achieve.. 

WHAT, WHY and HOW have therefore become a simple mantra I carry into every transformation. Start with WHAT you want to change and WHY it matters to the business. Only then decide HOW technology, data and AI can enable it. Too often, organisations start with the HOW and then look for a problem to solve. 

Another important lesson is that transformation cannot be something Data, AI or Technology teams do to the business. The business must own the outcome; technology must enable it. That shared ownership is what turns a technology programme into real transformation. 

Ultimately, success is not measured by how modern the platform is, how much data has been migrated or how many AI models have been deployed. It is measured by what has fundamentally changed, how much better, faster and more effectively the business now operates. 

Governance is often seen as something that slows innovation down and perceived as if you’re driving with hand brakes on. I see it very differently. Good governance gives you trusted data, the right security and access controls, and ultimately the confidence to innovate faster.

At HSBC, I learned that putting governance in after technology has scaled is difficult, expensive and often too late. Data quality, security, privacy and Responsible AI need to be built in from the start, not added as an afterthought. 

At the same time, governance has to be practical and proportionate. Not every dataset, AI model or use case carries the same level of risk. 

The future, in my view, is governance by design and controls embedded into the technology, with clear accountability, automated guardrails and continuous monitoring. 

Ultimately, governance should not tell an organisation how slowly it needs to move. It should give it the confidence to move fast, safely and at scale. 

This question is very close to my heart because I started my journey as an end-to-end solution and data architect over 15 years ago. It built a strong foundation in me and a belief I still carry today: strategic architecture cannot be designed only for today’s requirements; it must be ready for tomorrow’s change. 

Business priorities will change, technologies will evolve, and AI will create opportunities we cannot fully predict today. Architecture therefore needs to be modular, interoperable and, importantly, built around reusable capabilities rather than individual applications or use cases/initiatives. 

At HSBC and BCG, I saw first hand that building modern cloud, Data and AI platforms was important, but the real value came when those foundations could be reused across markets, functions and new use cases without rebuilding everything each time. 

At the same time, I have seen architecture become unnecessarily complex in pursuit of a perfect future state. Therefore the best architecture creates the right foundations today while giving the business freedom to evolve tomorrow. 

For me, the real meaning of scalability is not simply handling more data or compute, but how quickly the business can respond to the next business opportunity without having to rebuild its technology every time. 

For me, culture doesn’t change because leaders announce a new strategy or deploy a new technology. Culture changes when people experience clear results delivered by data and AI and see a better/improved way of working. 

One thing I have learned from leading large global teams and transformations is that technology adoption cannot be pushed from the technology function alone. Leaders need to create the right environment, make data easily discoverable, accessible and trusted. It doesn't stop here though, we should then give people the tools and skills to use it, and most importantly, demonstrate through their own decisions that data matters. 

The same applies to AI. We cannot simply give employees GenAI tools and call it adoption. People need to understand how AI can make their own work effective and efficient by

removing repetitive work or solving problems they could not solve before. When people see that value personally, adoption starts to become organic. 

I also encourage experimentation. Give teams the freedom to try, learn and even fail within clear guardrails. Some of the best innovation comes from people who fail fast and are closest to the business problem. 

Ultimately, you don’t create a data and AI culture by telling people to use data and AI. You create it by making it the easiest and most valuable way to get their work done. 

With AI, particularly agentic AI, we’re no donut entering into a different era where we will shift towards understanding context, reasoning across data, making recommendations and eventually executing parts of business processes. That changes the role of technology from simply supporting how a business operates to actively participating in how the business operates. 

But as I’ve said throughout this interview, this future will only work if the foundations are right. AI agents need trusted data, business context, semantic understanding, secure access and clear governance. In summary we need AI ready data 

With AI I also see that there will be no one clear owner of the business/IT processes and boundaries between data, applications and AI beginning to disappear. We will increasingly design intelligent business systems where these capabilities work together rather than as separate technology layers. 

For leaders, the opportunity is much bigger than deploying the next generation of AI tools. It is to reimagine how decisions are made, how work gets done and ultimately how the enterprise itself operates. 

That, for me, is where the next generation of enterprise transformation begins.