Patrick Thomas says the most durable skill in an AI-driven workplace is judgment written down, the discipline of turning tacit expertise into instructions an agent can execute. As Head of Communications, Consumer Apps Marketing at Google, he puts that belief into practice by codifying his own communications instincts into multi-agent systems that source and score content topics, draft in his voice, and route critical judgment calls back to human review.
With a 15-year career spanning global product policy, Google Cloud’s first brand narrative, and seven years leading EMEA marketing communications in London, Thomas treats AI adoption as an organizational and cultural challenge as much as a technical one.
Named to our Marketers to Watch list, Thomas shared his perspective on treating AI as an operating system rather than a content generator, why AI adoption spreads socially rather than through top-down rollouts, and why doing the work matters more than waiting for a title.
What book or podcast do you recommend to marketing leaders?
Goldratt’s Rules of Flow by Dr. Efrat Goldratt-Ashlag. It translates the core principles of bottlenecks directly into modern project management. As marketers and leaders increasingly orchestrate AI agents, we are discovering that managing agentic workflows is fundamentally an exercise in organizational science and operational flow.
The most actionable concept I use daily is “Full Kit”: assembling all necessary context, success definitions, constraints, and source materials before beginning execution. Our natural instinct is often to jump straight into work, but missing context forces systems (and humans) to pause, thrash, or make assumptions. Handing off a task to an AI agent without a Full Kit guarantees you will get back something useless. Assembling the Full Kit first is the difference between an agent that hallucinates noise and one that executes with precision.
How are you and your team currently using AI?
I treat AI as an operating system for how I work, not just a content generator. I’ve codified my own communications judgment into multi-agent systems: capturing my voice from real writing samples rather than abstract adjectives, preparing executive briefings, sourcing and scoring content topics, and automating repetitive synthesis while routing all critical judgment calls back to human review gates.
Across the organization, we teach this exact method through “AI Power Hour,” a program I spun up last year that continues to expand: start with a recurring friction point, map the actual workflow, separate what can be automated from what requires human judgment, and build the smallest useful system. We’ve found that whether you’re using AI for individual productivity or at org-scale, success requires building on a solid foundation of structured context.
What’s a prediction you have for marketing over the next few years?
Clever prompting will stop being a skill anyone cares about. The AI models available to consumers continue to get faster and more powerful. Tools like Gemini Spark, Grok Bot, and ChatGPT Work are bringing autonomous agents directly to non-technical professionals. In this world, the durable advantage comes from “judgment written down”: teams codifying how they decide, what excellence looks like, and where human review is non-negotiable into repeatable systems. Marketers who translate tacit expertise into instructions an agent can execute will set the pace, and they don’t need to be technical to do it.
Furthermore, AI adoption will remain stubbornly social. People don’t learn AI from top-down tool rollouts; they learn from watching a peer hand off a real piece of work. Organizations that treat AI adoption as a communications and cultural enablement challenge will move much faster than those treating it as a software procurement exercise.
What’s the most innovative or exciting project you’ve worked on recently?
Rebuilding my own role into a case study for how to significantly increase personal productivity with AI and agentic workflows. Over the past year, I’ve automated repeated tasks at org-scale and built systems that codify years of hard-won experience and communications instincts into skills that AI can use to get work done fast. To name a few examples, I rely on agents for proactive topic sourcing and scoring against individual expertise, for recommending novel angles that will resonate with readers, and for voice-calibrated first drafts that I can refine. Carefully crafted skills ensure consistent, high-quality execution every day. Google Antigravity featured this as its sole non-technical case study, proving that deep domain expertise matters more than an engineering background when building with AI.
I also write candidly about these experiments in cold start, my newsletter where I share copyable workflows and the exact breakdowns encountered along the way. The failures are where the actual learning happens.
What’s the most pressing business challenge you’ve faced in the last year and what have you done to solve it?
Moving an organization with hundreds of people from passive AI interest to active, daily practice. In our team, we recognized that AI was pivotal, but translating enthusiasm into practical daily action was tough. Many marketers felt caught between generic chatbot answers and tool fatigue.
We started by reframing this as an organizational communications challenge rather than a tech infrastructure problem. As mentioned, I founded AI Power Hour, a weekly org-wide session anchored in live, messy workflows, teaching people to become “business process engineers” and reimagining their existing professional judgment as structured steps. Power Hour has been a strong success, and I had the privilege of sharing the blueprint with our central Marketing Training team, who have scaled a version of the curriculum to thousands of marketers worldwide.
What leadership muscle is most important for marketers to exercise?
Delivering clarity. The most vital muscle right now is taking complex, ambiguous shifts — like AI transformation — and making them legible and actionable for your team. That requires diagnosis before prescription, showing your own work (including what failed), and setting clear boundaries between what machines can accelerate and where human judgment and empathy are most needed.
What’s the most game-changing piece of career advice you’ve ever received?
“Do the work before someone gives you the title.” Throughout my career, the greatest leaps came from spotting an unaddressed problem — whether building a missing regional narrative, launching a grassroots AI adoption curriculum, or automating away real operational friction — and demonstrating value through execution rather than waiting for a formal mandate.
What gives you energy and inspiration outside of work?
My family provides my core anchor and perspective. Beyond that, I get energy from helping people learn and work differently, whether by writing in public or building side projects that show what’s possible (like the AI tutor robot I designed and built for my kids ahead of back-to-school). I’m also getting back on the tennis court and coaching U8 soccer for the first time while pursuing a U.S. Soccer Federation grassroots coaching license. The common thread across all of them is the satisfaction of deliberate practice, curiosity, and getting slightly better at a craft each day.
Marketers to Watch is a recognition series to spotlight highly innovative and forward-thinking marketing leaders in the community. If you have someone you’d like to nominate for the series, apply here.
