The excitement around artificial intelligence is undeniable, but so is the sober reality: most initiatives get stuck in pilot mode or fail to deliver a meaningful return. The market is learning a hard lesson that winning with AI has less to do with the algorithm – and more to do with the foundational data that powers every organization.
In this first of our three-part series on AI, we unpack why a robust data strategy isn’t just a prerequisite for AI – it is the only way to succeed.
The productivity paradox: Why AI hasn't changed everything (yet)
It’s a story we keep hearing over and over these days: We're told AI will reinvent industries, create unprecedented efficiency – and unlock dozens of innovations. But for many leaders, the reality on the ground feels far less revolutionary.
Nobel laureate Paul Krugman recently compared AI to electricity in the early 1900s. The technology existed and was full of potential, but its impact on productivity took decades to materialize. The reason? Companies couldn't just plug a motor into their existing steam-powered factory: They had to fundamentally re-architect their workflows, factory layouts, and operating models to use its true power.
We are in a similar moment now. AI is a general-purpose technology, but its potential will only be unlocked by companies that do the foundational work of building their "wiring" – a modern, sophisticated data infrastructure.
The platform shift: Lessons from history
Every generation faces a platform shift that separates the innovators from the ones staying behind.
When electricity became viable, Northern textile mills in England, built vertically around steam power, saw the transition as too disruptive. New Southern mills, however, were built horizontally from the ground up, designed specifically for electric power. They were more efficient, lower cost, and faster. The North's hesitation cost them their market leadership.
A century later, a similar story unfolded. In the early 2000s: Palm, the undisputed leader in PDAs, hesitated when the mobile platform shifted to touchscreens and app ecosystems. By the time it adapted, the market had moved on. It wasn’t a lack of innovation that made these giants obsolete – it was a failure to re-architect their core operations during a critical platform shift.
The great disconnect: Why most AI initiatives fail
This brings us to the central paradox of AI today. While most organizations struggle, pioneering leaders are already integrating AI into their core operations to achieve tangible results:
- Walmart is leveraging AI for everything from complex supply chain optimization to dynamic, real-time pricing – resulting in an estimated $75 million annual savings from optimized logistics.
- Siemens and BMW are using AI on the factory floor for predictive maintenance and production line optimization – reducing unplanned machine downtime by up to 50%.
- AT&T is applying AI to manage network performance and proactively reduce customer churn at scale – cutting daily call analytics processing time with an open-source solution from 15 hours to less than 5 and lowering associated costs by approximately 35%.
- British Airways is utilizing AI-powered tools to analyze fleet and airport data, proactively reducing flight delays – achieving a record 86% on-time departure rate for flights from Heathrow.
But to understand this success: It’s not just about having better AI and machine learning algorithms – it's because these companies have spent years treating data as a strategic asset. They did the foundational work first.
Meanwhile, most organizations are stuck. Despite heavy investment, their AI projects stall or underperform. The reason for this disconnect isn't a technology gap; it's an organizational and strategic one, deeply tied to their business and data strategy:
- The strategy-tech divide: Business and tech teams often operate in silos. While business teams lack visibility into how the data works, technical teams focus on building optimal solutions – often without a clear grasp of the actual business need. The result? Technically impressive models that fail to deliver meaningful business impact.
- Inconsistent data assets: AI is greedy for data, but it needs to be the right data to ensure data quality: clean, structured, and well-documented. Most organizations run on a patchwork of legacy systems and integrated data silos, making it nearly impossible for AI to generate reliable or meaningful insights.
- Ambiguous governance: limited and ambiguous governance with no clear ownership and responsibility structure – leads to misaligned responsibilities and data models that fail to accurately reflect business priorities.
The fix: Building your data foundation
A successful AI program doesn't start with algorithms – it starts with a disciplined approach to building and managing data as a core business asset. Success depends on getting these AI foundations right. This includes …
- Clear, business-aligned use cases
- Data models tied to those use cases
- Automated, scalable implementations
- Defined ownership and governance
- A layer of intelligence that connects insights to action
While the principles are clear to many organizations, the reality of implementing them across siloed teams and legacy technology is the single biggest barrier to success. This is where our expert can help you bridge the gap between your ambition and your reality.
How Simon-Kucher Elevate builds your data foundation for AI business cases
At Simon-Kucher Elevate, we help organizations build these foundations through a pragmatic, step-by-step approach within a short timeframe that connects your data strategy to your commercial business outcomes:
- Start with commercial value: We work with your team to identify the most critical AI use cases within specific domains that, if answered, would drive the greatest commercial impact. These use cases become the North Star of your AI data strategy, ensuring every investment is aligned with what matters most.
- Model data around business processes: We design data models that reflect how your business actually runs and makes decisions, not just the constraints of your legacy IT architecture. This often requires you to develop a data model that rewires how data is collected and connected, a process we can deliver in as little as 3-5 weeks per business domain.
- Governance by design: We help you implement a lightweight but effective data governance framework where clear owners, stewards and custodians for critical data assets are defined. This ensures accountability and makes data quality a shared responsibility, not a bureaucratic exercise.
- Provide a clear implementation roadmap: To move beyond one-off projects, we provide a clear, actionable roadmap for using modern tools to automate your scalable data pipelines and model deployment, reducing friction and ensuring consistency.
Turning data into the ultimate fuel for growth
The electricity of the 21st century won't flow through old wiring: The companies that win in the age of AI won't be the ones that simply buy the latest technology. They will be the ones that do the hard, strategic work of building a robust data foundation first. If your organization waits, you won’t fail tomorrow – but you’ll slowly fall behind.
So, it’s time to take the steering wheel – and get control over your AI data strategy now. It’s time to join the tech revolution with Simon-Kucher Elevate!
Interested in more? In our next article, we'll debunk the most common AI myths and tackle the implementation challenges that derail most projects. Or just reach out to our Simon-Kucher Elevate experts to discuss your specific data and AI challenges.