Marketing AI Week in Review – May 4–10, 2025
This week offered a clear view of how quickly AI is moving from experimentation to execution. Big tech and major agencies are redefining brand, workflow, and productivity strategies around AI. Meanwhile, enterprise and industry players are grappling with infrastructure, data readiness, and the people side of change. In higher ed and healthcare, AI adoption is deepening—but not without debate. From agentic automation to open-source disruption, the themes this week are less about what’s possible and more about what’s already underway.
WPP Goes All-In on AI Workflow Automation
Last week I spoke at the Mirren Live conference in NYC, but also took in several fascinating talks that signal where agencies are headed with the use of AI. Brian Brown from Razorfish showed their amazing homegrown AI platform, a one-stop-shop for agencies seeking an AI-driven strategy and creative partner. Also last week came news that mega agency WPP is remaking its agency model with AI at the core. More than half its global workforce now uses its proprietary platform, WPP Open, featuring tools like Shower Thoughts (for early-stage creative prompts) and Brains (which uses LLMs and synthetic focus groups to test campaign ideas). This isn't a moonshot project—it's already happening. And for any agency still debating how seriously to take AI, WPP’s rollout is a clear signal: the transformation is operational, not theoretical.
OpenAI Reverses Course, Says it Will Remain a Nonprofit
One of the biggest pieces of news to come out of AI land last week was OpenAI’s announcement that it’s abandoning plans to transition from a non-profit to a for-profit operating model. If you remember the Sam Altman firing drama from November 2023, this was apparently one of the reasons for the firing. Public pressure has been mounting for a while for OpenAI to stay true to its original charter, and with this move, the AI giant is committing once again to better transparency, clearer governance, and more alignment with its mission to build safe, broadly beneficial AI. These changes come amid growing scrutiny of OpenAI’s influence and business model. For leaders watching the power dynamics in AI unfold, this is a reminder that how these tools are governed matters just as much as how they’re built. The structure behind the tech will shape who benefits—and who doesn’t—from AI’s rise.
Sam Altman Pushes for "Light-Touch" AI Regulation
In other OpenAI news, CEO Sam Altman and Rep. Jay Obernolte made a joint push last week for U.S. dominance in what they’re calling the “Intelligence Age.” Speaking at a Stanford event, both advocated for light-touch regulation to keep innovation moving—and to stay ahead of China. Altman emphasized that the U.S. must lead not just in technology, but in how AI is governed and deployed. For business and policy leaders, it’s a signal that the AI conversation is shifting from “what should we do?” to “how fast can we do it—and who gets there first?”
Microsoft Taps Meta’s Former Creative Lead to Rebrand Copilot
If you’ve spent much time using Microsoft Copilot, you already know that it’s…not quite there when it comes to user experience, usability, etc. Functions that seem like they should be connected, aren’t. Agent capability is both ubiquitous and non-existent. Similar tools are called different things. In short, it feels very much like a product being created in real time by a massive corporation. Microsoft is apparently aware of this, as they recently hired Mark D’Arcy, ex-Meta creative exec, to lead branding and UX for Copilot. His new role? Reimagining Copilot as more than a productivity assistant – it’s being positioned as a brand you can trust. The best way to build trust? Make the tool easy to use and have it actually do what you say it can do.
First-Party Data Is the Secret Sauce for Agentic AI
Agentic AI—tools that act independently to complete complex tasks—only delivers real value when fueled by high-quality, first-party data. That data gives AI agents the context they need to make accurate, personalized decisions and drive meaningful outcomes. Without it, automation risks becoming generic or error-prone. This article reinforces a critical point: enterprises chasing AI scale need to invest just as heavily in data infrastructure, governance, and enrichment as they do in the tools themselves. For marketers and IT leaders, it’s a wake-up call—your data isn’t just an asset, it’s your AI’s decision-making engine.
Mistral Launches Enterprise AI Suite With “Le Chat”
Mistral doesn’t get nearly as many headlines as some of the other AI companies, but the French company has quietly been building its capabilities and street cred. Now they’ve moved into the big time by debuting an enterprise-grade chatbot (Le Chat) and updated foundation model (Mixtral 8x22B), pushing deeper into the business AI market. Designed for transparency and control, this rollout is a big step toward commercializing open-source AI – and a signal that AI companies are increasingly seeing enterprise as the best path to more rapid AI adoption and bigger marketshare.
Can AI Be Too Nice? The Problem with Brown-Nosing Bots
A couple weeks ago, I wrote about how my MattGPT had become a passive-aggressive yes-man. This recent commentary suggests that as AI tools become more conversational, a new issue is emerging: they’re getting too agreeable. This piece explores how chatbots trained to avoid conflict or confusion can end up reinforcing user biases, sidestepping hard truths, or simply telling people what they want to hear. The result? Tools that feel helpful but lack intellectual rigor—especially dangerous in professional settings where accuracy matters. It’s a reminder that AI isn’t just about capability; it’s about calibration. If we want useful collaborators, not sycophants, we need to train AI to be honest, not just polite.
Chronicle: Professors Must Become “AI-Aware”
If you’ve spent much time in and around higher ed, you know that change comes slowly. Online education was decades old by the time the pandemic hit, yet the majority of U.S. universities had either never tried it or only had limited experience – leading to a massive scramble when online became the only option. AI adoption is seemingly on a similar trajectory, and a new piece in The Chronicle of Higher Education suggests that faculty can’t afford to ignore AI. As student usage explodes, institutions need clear plans for training educators on how to use—and teach with—AI. This isn’t just about academic integrity; it’s about preparing graduates for a world shaped by these tools.
Hospitals Are Embracing AI for Clinical Notes
At HMPS a couple weeks ago, there was one phrase I heard as much or more as any other: Ambient AI. Now, a national survey of 43 U.S. health systems shows growing adoption of generative AI tools for clinical documentation—particularly “ambient notes” that auto-generate summaries during patient visits. Clinicians report high satisfaction, especially in reducing burnout and improving efficiency. While AI is already well-established in imaging and radiology, ambient tools are emerging as a new frontline solution. Challenges remain—like cost, governance, and model maturity—but this is a clear signal: AI is moving from back-office experimentation to real-world impact in healthcare operations. For health system leaders, the opportunity is no longer theoretical—it’s operational.
AI is transforming customer service—but not by replacing humans. The real impact comes when AI supports agents behind the scenes: summarizing tickets, routing issues, or drafting responses. This frees up staff to handle the complex, emotional, or high-stakes moments that actually build customer loyalty. The article makes a clear case for a “human-first” model, where AI enhances—not eliminates—human connection. For marketing and CX leaders, the takeaway is simple: don’t automate empathy. Use AI to remove the noise so your people can focus on the moments that matter. That’s where real differentiation happens.
State of Martech 2025: AI Requires a New Playbook
The State of Martech 2025 report lays out a clear message: AI is changing everything—but only if your organization is ready to change with it. Marketing and ops leaders are being pushed to adopt more technical skillsets, prioritize open and interoperable platforms, and rethink how their stacks are managed. AI’s value lies in its ability to work across systems, unlock unstructured data, and drive continuous optimization. But with that comes complexity. The report encourages agile thinking, cross-functional alignment, and a build-to-adapt mentality. Martech success in the AI era won’t be about tools—it’ll be about orchestration.
Netflix Tests ChatGPT-Powered Search Feature on iOS
Netflix has quietly launched a new ChatGPT-powered search feature for iOS users, offering natural language recommendations and faster content discovery. Instead of typing titles or genres, users can ask questions like “What should I watch after Breaking Bad?” and receive personalized suggestions. The feature is currently in limited testing, with no word yet on Android or broader rollout. For marketers and product teams, this is another sign that AI-powered interfaces are moving from novelty to mainstream—reshaping how users search, choose, and engage with content in everyday apps.