Why you need an AI Strategy: 4 barriers preventing you from capturing value
Laura Stevens
Managing Director of Data & AI
While most companies are launching exciting AI pilots or integrating off-the-shelf AI tools, only some manage to scale their initiatives to create long-term business value. That’s because AI shouldn’t be treated as a plug-in but rather as a systemic shift that affects how organizations create value, make decisions, and structure teams.
Without a clear strategy, AI becomes another cost center, not a driver of growth. In this article, we outline four common barriers that block AI from scaling, and why a strategic approach is the only way through.
Why AI projects stall
Despite growing investments, most companies are still stuck in experimentation mode. According to Informatica (January, 2025), two thirds of businesses admit they haven’t moved most of their initiatives beyond pilot phases. Based on BCG reports (October, 2024), only 26% have developed the internal capabilities needed to generate real value from AI.
Often, organizations treat AI as a collection of tools rather than a coordinated capability. Every team is experimenting in isolation, and new AI-powered features are added to existing workflows instead of rethinking how work could be done.
To move beyond the illusion of progress, organizations need:
- Clear prioritization of which use cases deliver real business value
- Alignment across teams to avoid fragmented efforts
- Top-down ownership to shift from isolated pilots to enterprise transformation
Barrier 1: Strategic and organizational barriers
The biggest obstacle to scaling AI often isn’t the lack of technical skills but of a strategic direction. When companies don’t follow a clear direction for new AI projects, they fall into one of two traps:
- Chasing every new tool without lasting impact,
- or freezing in indecision while comparing the overabundance of potential solutions.
The real cost of not having a strategy
Scattered pilots that don’t connect to core business goals lead to duplicated efforts, inconsistent tools, rising costs, and long-term technical debt. Despite that, many companies assume that defining an AI strategy upfront is expensive or slows things down.
But in reality, the absence of strategy is what creates chaos:
- Teams build the same thing twice
- Leaders struggle to evaluate what’s working
- Infrastructure decisions get made ad hoc
A clear AI strategy provides decision-makers with:
- Clarity: Knowing on where to invest
- Coordination: Aligningment across business and technical teams
- Value: Reusing models, data, and tools instead of data silos
- Confidence: Knowing what works and stopping what doesn’t
Leadership fluency is a hidden bottleneck
Many executives feel pressure to act on AI but aren’t confident in what questions to ask or how to steer the initiatives. They struggle to separate hype from value and often expect instant results without understanding the trade-offs.
Lack of AI fluency across the executive team also often creates a mismatch between expectations and execution. A well-defined strategy helps close this gap by making roles clear, surfacing the right questions, and connecting technical work to tangible business outcomes
Rethinking ROI
Traditional ROI frameworks don’t always apply to AI. Leaders often expect the same clear business case they’d use for software development: defined cost with immediate return. The real value is hard to isolate, especially when AI augments human tasks rather than fully replaces them. Often, there’s also no baseline to compare against, particularly for first-time use cases.
As a result, AI initiatives may look unconvincing in a spreadsheet, even if they unlock significant long-term value. But this value tends to be compounded, indirect, and long-term.
When evaluating the ROI of AI initiatives, companies often fail to account for:
- Compounding value of automation across multiple workflows
- Reusability of AI components (models, pipelines, prompts) across teams
- Enablement of completely new capabilities, such as forecasting or real-time personalization
AI strategy can reframe the ROI conversation from “what do we save now” to “what capabilities are we enabling”, such as:
- Understanding the “strategic acceleration”, like faster decision-making or reduced dependency on specialized talent
- Measuring time-to-insight rather than time saved
- Tracking cross-functional reuse of data and tools
Barrier 2: Data and technical barriers
Most companies underestimate how much their existing data infrastructure can limit AI progress beyond initial pilots because the technical foundation isn’t built for scale.
AI makes technical debt visible
Scaling AI projects can amplify the existing technical debt the company has learned to live with.
For example, systems that weren’t designed to talk to each other now need to be orchestrated end-to-end, manually maintained spreadsheets can’t support model training at scale, and pipelines that were “good enough” for simple dashboards can’t handle the demand of real-time inference of AI models.
These are the most common blockers making even simple AI applications hard to implement:
- Data silos across departments
- Poor data quality and missing metadata
- Legacy IT systems that don’t integrate easily
- Fragile pipelines that break under load
- Over-reliance on centralized AI/IT teams, creating bottlenecks
What scalable organizations do differently
Organizations that scale AI successfully don’t wait for the perfect data. But they still build with the future in mind. Instead of treating data as a static resource, rethink how AI will change your ways of working in the long-run:
1. Build data foundations around specific use cases, not abstract data goals. Start by asking: “What outcome are we trying to achieve?” — then work backward to define the data you need.
2. Structure systems in a modular, flexible way so tools and data can connect easily. Avoid building isolated pipelines that can’t scale or be reused.
3. Give cross-functional teams hands-on access to data and the ability to improve it as they go. Don’t rely on centralized data owners or rigid governance processes.
Actionable AI strategy to drive efficiency and revenue growth
We set, stress-test and refine your AI strategy – with a tangible roadmap and business case based on what it takes to implement AI at scale and adopt AI-native operating models.
Barrier 3: Governance and operating model barriers
Many AI initiatives fail not because the tech doesn’t work but because the organization around it doesn’t adapt to the new ways of working. Even with the right data and a strong use case, execution breaks down when roles are unclear, teams are misaligned, and there’s no structured way to manage AI across the business.
The most common operating model issues include:
- No clear governance or accountability for AI initiatives
- Weak collaboration between business and tech teams
- Lack of executive understanding or ownership
- Siloed efforts or “competition” between departments
Scaling AI requires structural clarity
Without alignment on who’s responsible, how decisions are made, and how initiatives are prioritized, AI stays stuck in isolated experiments. Today, AI sits at the intersection of product, data, operations, and strategy. That makes cross-functional coordination essential.
Without organizational guiderails in place, even promising pilots stall in handover, and efforts get duplicated. Always keep in mind:
- Clarity on ownership: Who owns and approves what?
- Clear funding models: Is this IT’s budget or a business unit’s?
- Shared success metrics: How do we measure value across teams?
Barrier 4: People and cultural barriers
So far, most AI roadblocks have shown themselves to be human, not technical. Many companies still struggle to embed AI into the way people actually work. That’s because the biggest friction often comes from mindset, habits, and incentives.
Common people and cultural barriers include:
- Change resistance and organizational skepticism
- No shared language or clear narrative around AI
- Treating AI as a tool, not a workflow redesign
- Execution gaps between pilots and rollout
- Lack of incentives to experiment or adopt new processes
Tool thinking vs. system change
In many companies, AI is introduced as a tool layered onto existing workflows, in the form of a chatbot on the existing website or a copilot inside the current IDE. But unless those workflows evolve, AI ends up being underused or misapplied.
Real value comes not from adding AI to how people already work, but from redesigning how work happens with AI at the center.
That shift requires buy-in from not just the leadership, but the whole company. Teams need to understand how their roles will evolve with these new systems in place, how decisions will change, and how AI can support (and not just replace) their expertise.
Pilot fatigue is real
When teams see pilot after pilot with no real impact or integration into daily work, enthusiasm fades fast. Over time, this creates a culture of skepticism where AI is seen as just another top-down initiative that won’t stick.
Overcoming pilot fatigue requires:
- Clear communication about how AI will be used and what success looks like
- Early involvement of end users to shape the solution
- Incentives to test and adopt new ways of working
Fixing cultural blockers isn’t a one-off effort. It’s an ongoing investment in fluency, trust, and alignment.
Organizations that succeed treat this as a core part of their AI strategy, not an afterthought once the tech is built.
The key takeaway: AI Strategy first
AI isn’t a feature or an add-on, it’s a force multiplier. But without a clear strategy, most companies won’t get past isolated pilots or tactical wins.
The barriers aren’t just technical. They’re structural, cultural, and organizational. And they compound over time. Missing owners or an underfunded pilot turns into fragmented tools, rising costs, and lost momentum.
Companies that succeed with AI do things differently:
- They align on where AI creates value, and focus their efforts there.
- They fix foundational issues. Not just their data, but governance and workflows.
- They invest in change, building trust and fluency across teams.
- They design for scale from day one, treating AI not as an add-on, but as a new way of operating.
If you want AI to deliver real outcomes, start with strategy.
Let’s talk about about setting your AI Strategy.
Managing Director of Data & AI
Laura Stevens, PhD, is the Managing Director of Data & AI, bringing a unique blend of strategic vision, analytical expertise, and leadership acumen. With a background in neuropsychology, business consulting and organizational transformation, she has successfully navigated a career spanning academia, consulting, and industry leadership. As a former VP Data & AI in an international organization, Laura has led large-scale Data & AI teams covering data science, machine learning, data engineering, data governance, and visualization. She is passionate about leading organizations through their data & AI transformation.
laura.stevens@boardofinnovation.com
Keep learning
More resources on AI driven growth
The post Why you need an AI Strategy: 4 barriers preventing you from capturing value appeared first on BOI (Board of Innovation).