This article was based on the interview with From PegaWorld: Pega CTO Don Schuerman on AI ambitions versus reality by Greg Kihlström, AI and MarTech keynote speaker for The Agile Brand with Greg Kihlström podcast. Listen to the original episode here:
Let’s be honest with each other. The C-suite mandate for “more AI” has been handed down. For many of us in marketing leadership, this has resulted in a flurry of pilot projects, a growing pile of invoices for token usage, and a distinct lack of predictable, CFO-pleasing results. We are awash in potential but struggling to convert it into performance. This isn’t because the technology is failing us; it’s because we’re applying a transformative technology to processes that were built for a different era. It’s the classic case of having a revolutionary new tool but using it to bang the same old nails, just faster and more expensively.
The chasm between AI ambition and tangible returns is the defining challenge for enterprise leaders today. The organizations quietly canceling their AI projects are not the ones that lacked vision; they are the ones that lacked a new operating model to support it. Navigating this requires a unique perspective—one that understands both the intricacies of the code and the nuances of a campaign brief. It’s why the insights from Don Schuerman, who holds the rare dual role of CTO and Head of Marketing at Pega, are so relevant. He sits at the precise intersection of technological capability and marketing reality, and his perspective offers a clear-eyed roadmap for moving from disconnected experiments to a governed, agent-powered enterprise.
“The Org Chart is Showing”: Process Must Precede the Platform
There is a temptation with any new technology to believe the latest version or the most powerful model will be the silver bullet. We focus on the features, the demos, and the promise of what the tool can do, often overlooking the hard work of figuring out what we should do with it. The recent Pega and Savanta survey finding that 96% of organizations succeeding with agentic AI first rethought their processes should be a wake-up call. The real barrier isn’t a lack of understanding of the technology, but a reluctance to change how we work.
Schuerman argues that this internal work—redesigning workflows and team structures—is the true unlock for AI’s value, far more than any incremental model update.
“I think that more so than, you know, what, what does Opus 4.8 bring us over Opus 4.7? I think that’s the real unlock for AI value, not, you know, getting the latest version of the model.” – Don Schuerman
For marketing leaders, this means confronting the legacy of our own org charts. We’ve built departments around channels—the email team, the social team, the web team—and our processes reflect this siloed reality. A request for a new campaign often becomes a clumsy, sequential handoff from one specialized team to another. As Schuerman points out, this “request-dispatch” mode of operation simply doesn’t scale in an AI-driven world. Applying AI here just accelerates a broken system. The solution is to move toward smaller, agile teams that possess all the necessary skills and are given the autonomy to own a customer journey or campaign from end to end. When you empower a team like this, AI agents become powerful collaborators—assisting with content, pulling in data, and ensuring brand compliance—rather than just another piece of software in a clunky assembly line.
From Copywriter to Copy Architect: Elevating the Human Role
The narrative of AI often defaults to replacement, stoking fears that creative and strategic roles are on the chopping block. While some tasks will certainly be automated, the more sophisticated view is one of augmentation and scale. Leaders who successfully navigate this transition will be the ones who reframe the role of their human experts from individual executors to strategic enablers. Schuerman’s story of a copywriter on his own team provides a perfect, and frankly, encouraging, illustration.
“Now her impact has gone from, you know, the couple of documents and website pages she sees every month to almost every piece of copy that the company might generate. And that’s like, that’s empowering if you treat it that way.” – Don Schuerman
Initially resistant to generative AI, this copywriter realized the tool didn’t eliminate her job; it fundamentally changed it for the better. Her role shifted from simply writing good copy to scaling good copy across the entire organization. Her expertise is now codified into the prompts, markdown files, and guardrails that empower everyone in the company—from sales to HR—to generate high-quality, brand-compliant content. This is a profound shift. The value of an expert is no longer measured solely by their direct output but by their ability to multiply their expertise through technology. For marketing leaders, this is the blueprint for the future of our teams. We must identify our experts and help them transition from being players on the field to being the architects of the playbook that AI agents will help execute.
Know Your AI: The Sculptor and the Watchmaker
Not all AI is created equal, and the current hype cycle has led many to treat generative AI as a panacea. This is not only inefficient but, in the context of the enterprise, can be incredibly risky. To apply AI effectively, we need to be discerning about which type of AI is right for the job. Schuerman offers a brilliant metaphor to guide this thinking: the sculptor and the watchmaker. The sculptor is creative, free-flowing, and discovers form through exploration—this is generative AI. The watchmaker is precise, consistent, and follows exacting rules—this is the realm of traditional machine learning and statistical models. Good marketing needs both.
“At runtime, the AI that they’re using is far more watchmaker-like… it is always explainable… and those models have the ability to be continuously evolved in self-learning at a pace that… to retrain an LLM is really expensive.” – Don Schuerman
He explains how clients like Wells Fargo use the “sculptor” AI at design time to brainstorm potential offers and actions. This is the perfect use case for generative models: creative, low-risk ideation. However, when it comes to runtime—making billions of real-time decisions that determine which offer a customer sees in milliseconds—they rely on the “watchmaker.” These models are fast, massively scalable, and, crucially for a regulated industry, completely explainable. They can tell you precisely why a decision was made. Using a large language model for that kind of task would be like asking a sculptor to build a Swiss watch; it’s the wrong tool, and the results would be unpredictable and impossibly expensive. As leaders, our job is to orchestrate this collaboration, ensuring we’re leveraging the creative power of the sculptor without sacrificing the reliable precision of the watchmaker where it matters most.
Making AI Economics Predictable in a Federated World
Two of the biggest anxieties for leaders investing in AI are the unpredictable costs of token-based models and the persistent challenge of integrating disparate systems. The “token meter” has made forecasting budgets a nightmare, while the dream of a single, unified marketing “hub” has remained stubbornly out of reach. Schuerman suggests the market is maturing on both fronts, moving toward a future that is both more economically sound and architecturally realistic.
“This world doesn’t work in hubs anymore. This world is increasingly becoming federated… I think it becomes less about there being a single orchestration hub… But if I’ve got MCP [Multi-Channel Process] on my agents, I can reuse the knowledge that lives in an agent in a bunch of different places.” – Don Schuerman
The move away from charging per token to charging based on work completed is a sign of a maturing industry, one that is beginning to align its pricing with the value it creates. But the more profound strategic insight is the shift away from the monolithic hub-and-spoke model. The idea that we will get all our systems—Adobe, Salesforce, Pega, etc.—to play perfectly together under one roof has proven to be a fantasy. The federated model Schuerman describes is more practical and, ultimately, more powerful. Instead of a single orchestration hub, we can build specialized, intelligent agents—like a “brand code agent” that embodies all your brand guidelines—and make them available as a service across the enterprise. That agent can be called upon by your marketing automation platform, your sales enablement tool, and your HR onboarding process, ensuring consistency without forcing everyone into a single system. This is how we finally break down silos—not by demolishing them, but by building intelligent bridges between them.
Summary
The path from AI ambition to predictable reality is not paved with the latest model releases or a bigger technology budget. It is built on the far less glamorous, but infinitely more important, foundation of operational change. It requires us to redesign our internal processes to be more agile and customer-centric, breaking free from the constraints of our own org charts. It demands that we empower our human experts, transforming them from individual contributors into strategic architects who scale their knowledge through AI. Success also hinges on strategic discernment—knowing when to call upon the creative sculptor of generative AI and when to rely on the precise, explainable watchmaker of machine learning.
Ultimately, to remain agile in this rapidly evolving landscape, we as leaders must foster a culture of hands-on experimentation. As Schuerman notes, this is not a spectator sport; everyone needs to be playing with the tools. But he offers one final, crucial piece of advice: we must also have the discipline to step away. The spark of human creativity, strategic judgment, and authentic connection—the very things AI cannot replicate—is nurtured by turning off the screen and engaging with the world. That is the spark we need to lead our teams forward. As he wisely puts it, we have to keep that human element alive, and “that means turning the AI off every once in a while.”
