Marketers Must Build Their Own AI or Risk Falling Behind Fast
At Cannes Lions 2025, a standout conversation unfolded at Brand Innovators between Alysa Taylor, Chief Marketing Officer for Commercial Cloud & AI at Microsoft, and Liz Bacelar, Head of AI and Advanced Analytics at Under Armour. This was a strategic deep dive into how marketing leaders must evolve and not just to use AI, but to build it.
Marketers Must Shift from Users to Makers
“Personalization, automation, and governance” might sound like buzzwords, but in Taylor and Bacelar’s dialogue, they became tactical imperatives. Taylor emphasized a key point: AI is not one thing. With over 20,000 models in Microsoft’s Foundry platform, each serves different purposes with some tuned for healthcare, others for content creation. The takeaway? Marketers need to fine-tune models to fit specific outcomes rather than relying on generic answers from generalized data sets.
Bacelar made the distinction crystal clear: “When you think about the general models, you get everything—good and bad. But when you use focused datasets, you get much more useful answers.” This is where marketers need to rethink their toolkits and curate the data behind them.
From Campaign Journeys to Always-On Conversations
Taylor described a significant shift inside Microsoft: from pre-planned customer journeys to “automated, responsive workflows” that adapt based on real-time customer behavior. For example, a sovereign cloud announcement that would have taken months to prepare was fully executed in nine days, thanks to AI-generated content, modular GPTs, and Copilot tools.
Rather than replacing marketers, AI sped up production and liberated teams. “It just speeds up the cycle of time to get content into the market,” Taylor noted. That speed also enables marketers to be more responsive, personalized, and aligned with real-time business goals.
AI Agents and the Power of Triggers
Taylor shared how internal AI agents at Microsoft can now detect deal closures and automatically generate case studies—complete with written and visual content. This “triggered storytelling” closes the gap between business activity and marketing execution.
The real opportunity, she explained, is when AI agents know your audience. Imagine an agent summarizing Build announcements for a developer, or delivering Foundry updates tailored to a customer’s exact interests. “Agents in this world of personalization are going to get very, very good,” she predicted.
Governance Isn’t Optional—It’s Core to Trust
Bacelar challenged the room by raising what used to be taboo in marketing: legal and compliance risks. Taylor agreed, explaining that Microsoft’s approach to “trustworthy AI” centers on data privacy, bias detection, and legal guarantees like copyright protection.
“Your data is your data,” she emphasized, underscoring that Microsoft does not use customer data to train base models—an important stance for marketers navigating internal legal counsel and procurement.
Change Management Is the Hidden Obstacle
Despite the gains, adoption is not automatic. Taylor relayed a conversation with Chanel, whose merchandising agent cut processing time by 90%, and still struggled with internal usage. The reason? Workflow discomfort.
Bacelar observed that adoption is often met with resistance when AI is pushed on teams. Instead, organizations should enable those closest to the challenge to build the solution. “The people that built it are then the ultimate champions of it,” Taylor echoed, referencing a successful implementation at Canadian Tire.
This is more than a UX issue. It’s a call for a new leadership mindset—one that trains, empowers, and includes teams in the AI creation process.
What Comes Next
The most telling quote may have come near the end of the session. “It’s knowing the applied application,” Taylor said. In other words, theory isn’t enough. Marketers must understand how AI integrates into their daily work and that starts with community and transparency.
Both speakers encouraged marketing leaders to share their trials and wins—through roundtables, case studies, and peer learning. The AI CMO, they suggested, isn’t just smarter or faster. She’s more connected, more curious, and more willing to get her hands dirty.
Because in this new era of marketing, the winners won’t be those who watched the AI wave roll in. They’ll be the ones who helped build the surfboard.
