Why top-down AI strategy isn't enough without bottom-up innovation

For the past two years, pilots have dominated the enterprise AI conversation. Most have now proven AI’s potential and know it can deliver value. Yet, despite that progress, only 1% of business leaders believe their organisations have reached AI maturity, says McKinsey.

 

The era of AI pilots is over. It’s clear that now, the most pertinent challenge is turning isolated successes into organisation-wide transformation. And that’s where many businesses are starting to stall.

 

The real challenge now lies in where the next wave of AI opportunities actually exist.

 

Sure, executive teams can define ambition, allocate investment and set strategic direction. What they can’t do is identify the thousands of moments across the organisation where AI could genuinely transform workflows. Those opportunities don’t appear in strategy documents; they live with the people doing the work every day.

 

Arguably, therefore, success in AI adoption will belong to organisations that recognise a simple but highly valuable truth: top-down strategy will only create value when it’s matched by bottom-up innovation.

 

 

Strategy sets direction; employees find opportunities

Leaders have a distinct role to play in AI transformation. They define the strategic direction, commit investment and establish a clear vision for the organisation. That leadership is, inarguably, essential.

 

But bigger picture thinking alone is not enough. Finding immediate value in AI depends on identifying the right use cases, and those are rarely discovered in the boardroom. They emerge from the day-to-day realities and friction points for the people doing the work: repetitive tasks that consume hours, decisions delayed by disconnected information, knowledge that’s difficult to access, or manual processes that have simply become accepted as "the way we do things".

 

The people closest to those problems are also the people best placed to imagine how AI could solve them. This is the gap many organisations seem to underestimate. Leaders naturally think in terms of business strategy, while it’s the employees who understand the operational reality. When those two perspectives come together, AI stops being a collection of isolated experiments and starts becoming embedded in the way work gets done.

 

 

Governance should act as an accelerator

Many leaders still see a trade-off between innovation and control. In reality, the organisations making the greatest progress understand that governance and innovation are not competing priorities. Good governance is what allows innovation to scale.

 

Employees are already experimenting with AI, whether organisations have sanctioned it or not. This makes the important question not whether employees are using AI, but instead whether they’re doing so within a secure, trusted environment, or creating a patchwork of disconnected tools and practices that become impossible to govern.

 

The wise direction here is creating the conditions for responsible experimentation to flourish, not tighter control over experimentation. In practice, that means giving employees access to approved enterprise platforms, clear guidance on data and security, practical training, and governance that removes uncertainty instead of creating it.

 

Contrary to common belief, strong guardrails allow teams to move faster because they no longer have to invent their own rules or second-guess what’s acceptable. It’s only when leaders stop viewing governance as a barrier and start treating it as an enabler that good ideas can be scaled beyond the individual or team in which they were first identified.

 

 

AI transformation is a people challenge

In our experience, one of the clearest differences between organisations that scale AI and those that don’t is that they stop treating AI as a technology programme.

 

AI pilots do not fail because the technology isn’t capable – it’s usually very capable. They tend to fail because organisations struggle to change the way work is organised around the technology.

 

Deploying AI is relatively straightforward. But AI adoption? Well, that’s something else entirely.

 

Realising AI’s full value requires people to trust the technology, understand where it creates value and redesign workflows around it, rather than simply adding another tool into existing processes. That requires a shift in mindset as much as investment in technology.

 

The greatest impact comes when AI fundamentally changes how work is performed, how decisions are made, and how products or services are delivered, not from isolated productivity gains.

 

This is where leadership behaviour becomes decisive. Centralising and controlling every decision will not generate progress. What does generate progress is establishing a clear strategic direction while empowering business functions to identify opportunities, test ideas and demonstrate measurable value.

 

 

Rethinking the operating model, not just the tech stack

Perhaps the biggest mindset shift for leaders is to stop asking where AI can be deployed and start asking, "How would we design the business differently if AI existed from day one?"

 

That one question changes the conversation entirely, because instead of layering AI onto existing ways of working, organisations begin redesigning processes, decision-making and customer experiences around new capabilities.

 

Transformation starts with a clear North Star from leadership, but sustainable value is created through continuous experimentation and delivery across the organisation. The strongest outcomes come when strategic intent is connected with empowered teams that can identify, test, measure and refine practical use cases before scaling the most effective ideas securely and consistently.

 

Enterprise AI has, undoubtedly, entered a new chapter. The era of experimentation is giving way to the era of execution. The winners in this next phase will likely be defined by how effectively they embed AI into the way their organisation operates every single day.

 

---

 

Andriy Terlyha is Chief Delivery Officer and Partner at Intellias

 

Main image courtesy of iStockPhoto.com and imaginima