This article was based on the interview with LiveRamp’s Daniella Harkins on distinguishing AI hype and real innovation by Greg Kihlström, AI and MarTech keynote speaker for The Agile Brand with Greg Kihlström podcast. Listen to the original episode here:
The pressure on marketing leaders to “have an AI strategy” is palpable. It’s a recurring line item on board meeting agendas and a constant hum in the background of every planning session. The vendor landscape, of course, has responded in kind, with every platform suddenly rebranded as a revolutionary, AI-powered silver bullet. This creates a challenging environment for those of us tasked with navigating the digital frontier. The risk isn’t just about choosing the wrong tool; it’s about being paralyzed by the sheer volume of noise or, conversely, being pushed into rushed, reactive decisions that serve the trend more than the business.
Distinguishing genuine innovation from well-packaged hype requires a clear-eyed approach grounded in business fundamentals. It’s a moment that calls for leadership, not just adoption. In a recent conversation, Daniella Harkins, SVP of Product Go-to-Market at LiveRamp, offered a refreshingly pragmatic perspective. Her insights cut through the speculative promises to focus on what matters most: solving real business problems, building on a solid data foundation, and empowering teams to learn in an era where the playbook is still being written. For marketing leaders looking to move beyond theory to tangible application, her guidance provides a practical framework for building an authentic and effective approach to AI.
Forget the Perfect AI Strategy; Focus on the Problem
One of the greatest sources of anxiety for leaders is the expectation of having a flawless, multi-year AI roadmap. In a field evolving at this pace, that’s not just unrealistic—it’s counterproductive. Chasing a perfect, all-encompassing strategy can lead to inaction, as teams become overwhelmed by the possibilities and the fear of making a wrong move. Harkins argues for a more grounded, iterative approach that begins not with the technology, but with the business itself.
“If anybody thinks that they are gonna have a perfect AI strategy today… it’s too early. It’s too premature. And if you think you do, chances are you don’t, and it’s going to change dramatically. And so instead of thinking about it across a multi-year strategy… I’m always, I say, ‘Listen, first and foremost, think about the business challenges that you’re trying to solve. Don’t get lost in the technology speak.’ Everybody wants to jump right to the how and right to the technology, but instead, I would say, what is it that we’re trying to solve for? Stay focused on that and pick one or two use cases that you can then start to test into.”
This is a critical mindset shift. Instead of asking “What is our AI strategy?” the more effective question is “What are our most pressing business challenges, and how might AI help us solve one of them?” This reframing immediately makes the task more manageable. It transforms a nebulous technological mandate into a series of discrete, value-driven experiments. Are you struggling with media optimization? Is campaign personalization falling flat? Are your operational workflows bogged down in repetitive tasks? Pick one. By focusing on a specific use case, you create a contained environment for learning. This allows your team to test, fail, and iterate without betting the entire marketing budget on an unproven, grand-scale implementation. It also provides a clear benchmark for success, moving the conversation from technological capabilities to measurable business outcomes.
Your AI Is Only as Good as Your Data Foundation
The well-worn adage “garbage in, garbage out” has never been more relevant. In the rush to implement sophisticated AI models and decisioning engines, the foundational work of data management can be overlooked. Yet, without a clean, connected, and ethically governed data infrastructure, AI is not a solution—it’s a multiplier of existing problems. An AI tool might help you make decisions faster, but if it’s working with fragmented or inaccurate data, it will only help you make the wrong decisions with greater speed and efficiency. Harkins was unequivocal on this point, pushing back against a particularly dangerous narrative gaining traction in the industry.
“I am hearing over and over again in the industry that identity just doesn’t matter as much anymore. And I actually think that that is a false narrative, because I think it matters more than any time in the past. I think it’s more critical to running our businesses than it’s ever been… I think if you walk into it saying, ‘I’m just gonna do what’s good enough, I’m gonna do this because it’s easy, and I’m not gonna worry about really building an identity strategy, and then connecting all of my data to that identity strategy,’… I think that is a blind spot, and I think it is something that will hurt organizations longer term.”
For enterprise marketing leaders, this is a call to action. The allure of the “Easy Enough” button—relying solely on an email address or a single identifier—is a trap. Today’s customers are not single-channel, single-device entities. They interact with brands across a complex web of touchpoints. A robust identity strategy that can responsibly resolve these fragmented signals into a coherent view of the customer isn’t just a prerequisite for effective personalization; it’s fundamental to maintaining consumer trust. Before you invest heavily in the next AI-driven personalization engine, take a hard look at your data foundation. Is your customer data unified? Is your identity resolution strategy sound? Do you have clear governance on the ethical use of that data? Answering these questions isn’t the exciting part of the AI journey, but it is, without question, the most important.
Embrace the Experiment: Standardization Isn’t Here, So Let Your Teams Learn
In most enterprise environments, efficiency is king. Redundancy is a dirty word, and standardization is the holy grail. When it comes to AI, however, applying this traditional mindset too rigidly can stifle the very innovation you’re trying to foster. The tools are new, the best practices are still emerging, and frankly, nobody has all the answers. In this nascent stage, learning is more valuable than perfect efficiency. Harkins encourages leaders to accept a bit of organizational messiness and give their teams the space to experiment, even if it means some duplicated effort.
“Nothing is standard yet. Nothing is standard yet. And be okay with people duplicating work… We’re going to be learning together. We’re potentially going to be doing work that somebody else has done, but it’s so important for us all to test it out, to get our hands dirty, and to not worry about what other people have done, but now start to learn that from other people. And so the idea of standardization, we’re not there yet. Not even close.”
This is liberating advice for leaders and teams alike. It grants permission to be imperfect. Instead of waiting for a central authority to hand down the “official” AI tool or process, this approach empowers individuals and teams to be proactive. Encourage your team members to try different tools for different tasks. One person might experiment with Gemini for research synthesis, while another builds a simple agent to automate reporting. Will there be overlap? Yes. Will some experiments fail? Absolutely. But the collective knowledge and capability your organization gains in the process is invaluable. This hands-on experience demystifies the technology and builds a culture of curiosity and resilience. The goal right now isn’t to find the one right way to do AI; it’s to build an organization full of people who are comfortable finding many different ways to solve problems with it.
The path forward with AI is not a straight line, nor is it a singular, top-down mandate. As Harkins’ insights reveal, the most effective approach is iterative, grounded, and human-centric. It begins by trading the pursuit of a perfect, all-encompassing strategy for a focus on solving specific, tangible business problems. This immediately anchors the work in value creation and provides a clear lens through which to evaluate the endless stream of technological possibilities. From there, it requires an honest and rigorous assessment of your data foundation, recognizing that a robust and ethical identity strategy is the non-negotiable bedrock upon which all successful AI initiatives are built.
Ultimately, this new era calls for a shift in leadership style. With so much of the tactical execution of marketing set to be automated, the role of the senior leader becomes less about managing the how and more about defining the why and the what. As Harkins puts it, this is an opportunity to move “from tactics to strategy,” freeing up valuable time to think, learn, and guide the business. By empowering our teams to experiment, by fostering a culture that values learning over perfect efficiency, and by keeping our focus squarely on the customer and the business, we can navigate the hype and harness the true promise of AI to build smarter, more agile, and more meaningful marketing organizations.
