This article was based on the interview with TransUnion’s Matt Spiegel on the AI confidence-readiness paradox by Greg Kihlström, MarTech and AI Adoption futurist and keynote speaker for The Agile Brand with Greg Kihlström podcast. Listen to the original episode here:
There’s a peculiar tension in the enterprise marketing world right now. On one hand, the C-suite mandate is clear: adopt AI, scale it, and prove its value. The confidence is palpable. In boardrooms and budget meetings, AI is discussed as an inevitability, a foregone conclusion on the path to greater efficiency and market dominance. Nearly every marketing leader will tell you they are well on their journey, with pilots spinning up across creative, analytics, and media activation. Yet, beneath this veneer of confident adoption lies a more complicated reality—a significant gap between our ambition for AI and our organizational readiness to actually harness its power.
This is the heart of the AI confidence-readiness paradox. Recent research from TransUnion highlights this disconnect: while AI adoption is nearly universal, a mere 13% of marketing leaders report having fully scaled it. This isn’t a story about failed technology. It’s a story about what happens when a powerful new capability collides with the stubborn realities of legacy data structures, siloed processes, and entrenched human behaviors. AI is proving to be a rather unforgiving magnifying glass, exposing the pre-existing operational cracks we’ve managed to work around for years. The journey from a promising pilot to enterprise-wide integration isn’t a technology problem; it’s an organizational one. And as many of us are discovering, it’s where the real work begins.
The Three-Headed Dragon of AI Stagnation
The allure of a pilot program is its isolation. In a controlled environment, with a dedicated team and a clean dataset, it’s relatively easy to demonstrate the potential of an AI tool. The trouble starts when that tool has to survive in the wild—the complex, messy ecosystem of a large enterprise. The reasons for stalling are rarely technical; they are deeply rooted in the organizational fabric. As Matt Spiegel notes, the barriers are predictable, but that doesn’t make them any less formidable.
“There’s really three core challenges that I’d identify. Let’s talk about data, let’s talk about processes, and let’s talk about people… The AI is only as good as the data you feed it… if you haven’t redone your org process and the processes around how you go to market, well, then you probably also have some friction that you can’t yet get there… And then the third piece of that is just the change model that goes with it, which is we as people, it takes work to get us to change and do new things.”
For marketing leaders, this is a familiar refrain, but AI raises the stakes exponentially. First, the data: a fragmented view of the customer, with CRM data in one system, e-commerce transactions in another, and media exposure data somewhere else entirely, is no longer just an inconvenience. When you feed this piecemeal information into an LLM or a predictive model, you aren’t just getting incomplete insights; you are actively training the system on a flawed reality. The model’s outputs may look intelligent, but they are built on a foundation of sand.
Second, the processes. Our organizations were largely built for a pre-AI world, structured around functional specialties like search, social, and analytics. Yet, the most powerful insights AI can deliver are inherently cross-functional. An insight about customer churn might originate from media interaction data but have profound implications for the CRM and on-site personalization teams. If your organizational chart and workflows still resemble a series of disconnected silos, those insights will die on the vine. Finally, the people. The challenge isn’t about convincing teams that AI is important; it’s about navigating the fundamental change to their daily work, a process that requires a deliberate and sustained change management effort.
Beyond “Clean Data”: The Imperative of a Unified Identity
The phrase “garbage in, garbage out” has become a cliché in the data world, but it’s never been more true. While most leaders acknowledge the need for clean data, Spiegel argues we need to think more structurally. The real foundation for enterprise AI isn’t just tidy data; it’s a coherent, unified view of the customer, anchored by a common identifier that persists across every touchpoint.
“I really do think that the core of the future of this people-based marketing world we wanna live in does start with a common view of identity, a common key that allows you to link different touchpoints into a common view, meaning you don’t want Matt and Matthew to be two different people in your database.”
This concept of a “common key” is the linchpin for moving beyond surface-level AI applications. It’s the mechanism that connects a user’s ad exposure on a publisher site to their activity on your website, and ultimately to their purchase in your CRM. Without it, you are essentially asking your AI to analyze a room full of strangers with no context. With it, you can build a rich, longitudinal view of a person, enabling far more sophisticated modeling, personalization, and measurement.
Achieving this unified view is, of course, easier said than done. It requires a strategic partnership between marketing and technology teams, and often involves leveraging partners and cloud data platforms to do the heavy lifting of identity resolution without having to move massive datasets. The critical takeaway for leaders is that this isn’t a task to be delegated and forgotten. Establishing a common identity framework is a strategic imperative. It’s the infrastructure that ensures your AI investments are being fed the high-quality fuel they need to generate real business value, not just interesting but ultimately unactionable outputs.
The Measurement Mandate: Owning Your Intelligence
With boards and CEOs demanding tangible ROI on AI investments, the pressure to prove value has never been higher. We’ve moved rapidly past the initial phase of “look at this cool thing it can do” to the far more difficult phase of “what did this cool thing actually do for the business?” The temptation is to rely on the metrics provided by the platforms and vendors themselves, but Spiegel offers a strong caution against this approach.
“One of the other things that I see happens a lot… is that marketers end up relying on the metrics from the media companies or their technology vendor aggregators. And what happens is, is everyone then claims too much credit… The best case scenario is you actually are extracting it from one channel, and you’re actually measuring across channels… the more that you begin to wanna internalize that intelligence in a cross-platform way, the more that you’re going to learn.”
This is a call for marketing leaders to take back control of their measurement narrative. Relying on siloed, self-reported metrics from individual channels or walled gardens is like asking each player on a sports team to grade their own performance—everyone will claim to be the MVP. True understanding comes from creating an independent, cross-channel view that you, the marketer, own. This allows you to see the complex interplay between different tactics and to attribute success more accurately.
Building this capability is the foundation for a credible AI measurement framework. It allows you to move beyond proxy metrics like “time saved” or “cost cut” and start answering the questions that matter: Did this AI-driven audience strategy lead to a higher conversion rate? Did our AI-optimized creative actually drive incremental sales? By internalizing this intelligence, you not only get smarter about your marketing mix but also build the business case for further investment. It shifts the conversation from a leap of faith to a data-driven strategy for growth.
The path from AI pilot to scaled enterprise value is littered with challenges, but they are not insurmountable. The confidence-readiness paradox isn’t an indictment of AI’s potential; it’s a clear-eyed assessment of the foundational work required to unlock it. The technology, for the most part, is ready. The question is, are we? The insights from experts like Matt Spiegel serve as a crucial reminder that our focus must shift from the tool itself to the organizational ecosystem in which it operates.
The leaders who succeed in this new era will not be those who chased the most trends or ran the most pilots. They will be the ones who had the discipline to address the unglamorous but essential work of unifying their data, re-engineering their processes, and building a culture of measurement and change. AI is an accelerant; it will amplify our strengths and our weaknesses with equal force. The organizations that embrace this reality and invest in building a solid foundation will be the ones that see their confidence translate into real, defensible, and market-leading results.