Why pragmatism in AI is driving the future of data strategies

GenAI has democratised AI adoption, offering easy entry points with pre-built models and APIs. While these tools address basic needs, enterprise-scale challenges demand more. Bespoke models and advanced machine learning (ML) capabilities require significant investment in infrastructure, data readiness, and technical maturity. This isn’t just about deploying flashy tools; it’s about ensuring scalability, reliability, and ROI.

Why building on a sand has never been a good idea

At the heart of AI success lies data and platform engineering, the unsung heroes of the digital transformation journey. Without data architectures leveraging all the benefits of cloud infrastructure, well-engineered pipelines, clean datasets, optimised storage solutions, and strong quality assurance, even the most advanced AI models fail to scale or deliver value.

Data and platform engineering is not a supporting act—it is foundational.

For instance, consider the implementation of real-time AI-driven recommendations. Behind the scenes, this requires real-time data pipelines to capture and process events, data governance to ensure compliance, quality, and consistency and a scalable architecture which leverages cloud technologies to handle peak loads and keep costs under control.

The power of fusion of data and software engineering

(and a handful of other disciplines)

Today, software and data are completely intertwined.

Modern platforms rely on seamless integration between data systems and software applications. This integration demands cross-functional teams capable of bridging gaps between software and data and delivering solutions end-to-end.

Predictive analytics is a good example of this shift. As organisations mature their data strategies, they’re moving beyond static dashboards to dynamic, real-time insights. This evolution requires automated data workflows to reduce manual intervention, real-time data syncing to feed predictive models and integration into decision-making systems so AI insights can directly inform operations or end-users – user experience and service design are more vital than ever, at least if at least one of your process touch points involves good old humans.

The result? Tangible value delivered at speed and scale, with businesses equipped to act on insights as they arise.

The data science is dead, long live data science

(and ROI)

Data science, once a crown jewel of innovation, is increasingly becoming commoditised.

With the rise and democratisation of GenAI and machine learning solutions, many use cases that historically would require a significant investment and an expensive in-house (or even more expensive if outsourced to a consultancy) Data Science capability, can be now implemented in a matter of weeks if not days. Being able to recognise what use cases are good candidates for off-the-shelf APIs and which ones require bespoke models and complete machine learning enterprise eco-systems is becoming more critical than ever – the difference in the cost of implementation can be more than 100-fold!

In other words, does every company under the sun need to train their own LLMs? Probably not. Can most companies benefit hugely from AI capabilities available as a service? Absolutely.

The broader market echos this shift to pragmatism. Investors, boards and senior management are prioritising ROI-driven, sustainable growth over speculative ventures which risk burning budgets and careers. AI is no longer a novelty, it’s a strategic enabler for achieving measurable outcomes.

Pragmatism also extends to talent. Instead of hiring large teams of data scientists, many organisations are now investing in machine learning engineers and platform specialists who can deliver both infrastructure and insight. This shift reflects the growing need for talent that can understand the theoretical but also the practical aspects of AI implementation.

Consulting in a changing landscape

The role of the consultant is also evolving. Gone are the days when presales engagements are pitches of what the tech can do, today’s customers increasingly expect strategic value upfront…clear guidance on framing objectives, building roadmaps, and demonstrating ROI.

Winning in this landscape requires blending advisory and delivery services seamlessly and more importantly…listening to your customer as there is no goodness without business insight.

Helping clients navigate the intertwined nature of software and data in the simplest way possible is essential, too many human logic connections just confuses and the need for simplicity of thinking is winning. It’s not enough to recommend a data strategy; consultants must also consider how the AI models integrate into existing software landscapes to ensure the end-user experience aligns with business goals.

The opportunity ahead

For businesses, the path forward is clear: embrace a strategic approach that combines bold AI ambitions with pragmatic execution.

Focus on foundational capabilities, invest in scalable data solutions, and prioritise outcomes over blind experimentation as the goal. Be creative in thinking about the use case, operational or strategic, they are out there, start simple and prove them.

The winners will be those who not only adopt AI but embrace how engineering, data, and strategy co-exist. Putting it simply, how an organisation drives AI ambition through creative use cases, delivery and measured ROI will likely win the early skirmishes in the age of AI.