By David Fouse
In our previous article, I shared how Pinkston is approaching AI implementation as fundamentally a communication challenge. Today, I want to explore another critical insight from our journey: the scarcity of experienced guides in this rapidly evolving landscape.
Seeking Expertise in Uncharted Territory
Over the course of our AI journey, we’ve consulted with numerous technology companies, including one of the major LLM players. These conversations were initiated with straightforward objectives: to describe what we see as our business challenges, explain where we are in our implementation process, and understand how their expertise might accelerate our progress.
What we discovered was unexpected. Despite speaking with genuinely knowledgeable technologists, we found ourselves in a peculiar situation—there are remarkably few “doctors” with the experience to help diagnose our specific organizational needs and prescribe a clear path forward.
This isn’t to disparage the capable professionals we’ve engaged with. Rather, it highlights a universal challenge: AI is evolving so rapidly that even experts are learning in real time. As Sarah Tavel of Benchmark Capital aptly noted in an Every podcast interview, “We’re all reinventing the wheel. [But] this is not how technology is going to get populated.”
The Advisor Becoming the Advised
The irony of this situation isn’t lost on us. At Pinkston, a significant part of our value proposition is acting as strategic advisors. When clients come to us, they often have an unclear understanding of what they need and how we might help them. Our expertise lies in asking insightful questions, defining the real issues behind their stated challenges, and developing strategic solutions.
It’s easy to understand this dynamic through the familiar example of visiting a doctor. We arrive with symptoms, expecting the doctor to ask insightful questions, identify the underlying issue, and outline a clear path to recovery.
Yet in our AI implementation journey, we’ve found ourselves in our clients’ shoes—seeking that same level of diagnostic expertise and finding fellow explorers rather than experienced guides.
Navigating the Fog of Innovation
This advisory gap occurs during what we might call the “fog of innovation”—that early stage of technological disruption characterized by excessive noise but limited clarity. In this environment, there’s a persistent feeling of tardiness, creating pressure to act without sufficient knowledge, which often leads to hasty implementations that fail to address organizational needs sustainably.
This experience is not unique to Pinkston. According to the IBM Institute for Business Value’s 2025 CEO Study, over the past three years, only 25% of AI initiatives have delivered expected ROI, and a mere 16% have scaled enterprise-wide. These statistics underscore the widespread challenge organizations face in moving beyond pilot programs to meaningful implementation.
High-profile announcements from companies about their AI initiatives provide tantalizing glimpses of possibility but typically without the detailed roadmap that would make their experiences replicable for organizations with different needs and constraints.
The DIY Imperative
Perhaps the most important lesson we’re learning is that AI implementation isn’t a responsibility you can fully delegate to external experts. You must understand where and how it will impact your specific business and then do the hard work of addressing the implications internally.
This realization has led us to develop a more self-reliant approach. Rather than waiting for the perfect advisor or comprehensive implementation playbook to emerge, we’re becoming more intentional about building internal expertise while continuing to learn from external sources.
Our Layered Approach to Implementation
As a strategic communication agency with multiple integrated divisions, our AI needs have both specific requirements that serve individual units and enterprise-wide considerations that work cohesively across the organization.
This complexity has led us to develop a layered approach:
- Build an Enterprise Foundation: Core AI capabilities and governance principles that serve the entire organization.
- Develop Divisional Solutions: Specialized AI tools and workflows tailored to the unique needs of each service area.
- Implement Cross-Functional Integration: Systems to ensure divisional AI initiatives enhance rather than fragment our collaborative work.
This approach aligns with findings from the IBM study, which identified “organizational silos/lack of collaboration” as the top barrier to innovation in AI implementation. Breaking down these silos is essential for meaningful progress.
Where We Stand Today
Our recent internal survey shows promising signs that our approach is working, though we’re far from declaring victory. While 58% of our staff agree that our AI rollout over the past three months has been successful, around 30% remain neutral. More encouragingly, 74% report feeling comfortable with our current implementation of AI, with only 6% expressing discomfort.
Perhaps most telling is the finding that approximately 20% of our employees don’t believe Pinkston has articulated a clear AI strategy. This disconnect—where comfort with the technology outpaces understanding of its strategic direction—signals an important communication challenge. While we’ve made technical progress, we still need to better articulate our vision and roadmap. This experience mirrors what the IBM study revealed: for many organizations, the technical aspects of AI implementation often advance more rapidly than the necessary cultural alignment and organizational understanding.
Lessons for Fellow Travelers
For organizations at similar points in their AI journey, we offer these insights from our experience:
- Embrace the advisor-explorer duality: The most valuable external partners may not have all the answers, but will join you in asking the right questions.
- Develop internal expertise: Invest in upskilling team members who can translate between your business needs and evolving AI capabilities.
- Prioritize use cases over technology: Focus first on specific business challenges rather than implementing AI for its own sake.
- Build feedback mechanisms: Create structured ways to evaluate what’s working and what isn’t across your organization.
- Foundation in data: As 72% of CEOs in the IBM study note, proprietary data is key to unlocking the value of generative AI. Ensure your implementation is business-driven and centered on your organization’s unique data and needs.
Most importantly, maintain transparent communication throughout the process. When the technology itself is evolving daily and expertise is in short supply, honest dialogue about expectations, limitations and progress becomes your most reliable guide.
Moving Forward
As we continue on this journey, we’re working to engage more of our staff in actively exploring how AI might benefit their daily work. Our survey shows progress, but implementing AI isn’t a one-time project—it’s an ongoing evolution.
The IBM study projects that 31% of the workforce will require retraining or reskilling within three years due to AI implementation. At Pinkston, we’re prioritizing engagement and, ultimately, education to ensure our team is prepared for this transformation.
What has your experience been with finding AI implementation guidance? Have you encountered similar gaps between technological promises and practical roadmaps? I’d be interested to hear how other executives are charting their own courses through these uncharted waters.
David Fouse is a Partner at Pinkston, a strategic communications firm based in the Washington, D.C. area.
