On Google’s Frontier CMO series, host Josh Bania sits down with JPMorgan Chase CMO Carla Hassan to unpack what it actually means to operate under pressure in modern marketing. The conversation spans trust, AI, and organizational design—but the most strategically important shift isn’t about technology. It’s about how AI forces a full rewrite of how marketing organizations execute.

This is not a tooling story. It’s an operating model story.

AI Doesn’t Upgrade Execution—It Exposes It

Hassan makes a point that most teams still underestimate: “Putting AI on a crappy process is a crappy process with AI.”

That line reframes the entire AI conversation.

Most organizations approach AI as an overlay—something you plug into existing workflows to accelerate output. But that assumption breaks quickly. AI doesn’t just speed things up; it removes the buffer that used to hide inefficiencies. Broken handoffs, redundant approvals, unclear ownership—these don’t get fixed by AI. They get amplified.

At JPMorgan Chase, the response wasn’t incremental optimization. It was a full teardown.

Instead of asking “Where can AI help?”, the team mapped execution end-to-end: from strategy to content creation to audience targeting to measurement. Only after that did they ask a more fundamental question—where should humans actually be involved?

That’s a very different starting point.

The OPERATE Lens: Execution Is the Constraint

This shift sits squarely in the Execution pillar of the OPERATE framework—how plans turn into action, how workflows actually function, and how teams deliver outcomes consistently.

Execution is where most AI strategies quietly fail.

Not because the tools don’t work, but because the underlying system wasn’t designed for speed, iteration, or modularity. Traditional marketing orgs are built around sequential workflows: strategy → creative → media → analytics. Each step introduces delay, translation loss, and dependency risk.

AI collapses those timelines. But if your execution model still assumes linear handoffs, you don’t gain speed—you create friction at scale.

Hassan’s team recognized this early. Instead of inserting AI into legacy workflows, they used AI as a forcing function to redesign execution itself.

That’s the move.

From Writers to Editors: Redefining Work at the Role Level

One of the clearest examples is how content production roles evolved.

Historically, SEO and content teams were built around writers—people responsible for producing articles at scale. With AI, that role shifts. The output is no longer constrained by writing capacity; it’s constrained by judgment.

So the role becomes editorial.

Teams move from generating content to shaping it—deciding what should exist, how it should be framed, and how it aligns with brand voice and intent. The result is both higher throughput and higher leverage. Hassan notes a 30% efficiency gain and a doubling of content output, but more importantly, the work itself becomes more strategic.

That’s not automation replacing labor. That’s execution being redefined at the task level.

Change Management Is the Real Bottleneck

The harder problem isn’t technical—it’s human.

Hassan is explicit about this: the biggest challenge is getting high-performing teams to believe that the new way is better for them, not just for the business.

That’s a non-trivial shift.

Most marketing teams are full of people who succeeded under the old model. Their identity is tied to how work gets done. Changing execution isn’t just a process update—it’s a status reset. Who creates value? Who makes decisions? Who owns outcomes?

If you don’t manage that transition deliberately, AI adoption stalls—not because of resistance to technology, but because of ambiguity around roles.

JPMorgan’s approach—testing new workflows against old ones—creates a bridge. It doesn’t force blind adoption; it proves superiority through performance. That’s a subtle but important design choice.

Speed Changes the Feedback Loop

Another implication of this execution shift is how feedback works.

In a traditional model, campaigns are tested over months. In Hassan’s model, content is deployed and evaluated in real time. Platforms like YouTube and social become live testing environments where performance data is immediate and actionable.

That compresses the learning cycle.

But it also requires a different execution discipline. You can’t rely on pre-launch validation when the market itself becomes the test. Teams need to be comfortable shipping, reading signals quickly, and iterating without over-structuring the process.

Execution becomes less about planning and more about responsiveness.

AI Forces Organizational Rebundling

As execution speeds up, the traditional separation between functions starts to break down.

Hassan points to an emerging trend: the rebundling of creative, media, and strategy into more integrated workflows. When output can be generated and tested instantly, the cost of handoffs becomes too high.

This isn’t just a structural change—it’s an execution necessity.

The future model looks less like specialized silos and more like integrated pods, supported by centralized, AI-enabled systems. The implication is clear: execution is no longer about managing complexity. It’s about removing it.

The Founder Takeaway: Redesign Before You Automate

The instinct to adopt AI quickly is correct. But the sequencing is usually wrong.

Most teams start with tools. The better approach is to start with execution design.

Map your workflows. Identify where decisions are made, where delays occur, and where ownership is unclear. Then decide what should be automated, what should be augmented, and what should be eliminated entirely.

AI is not the upgrade. It’s the audit.

The companies that win won’t be the ones with the best tools. They’ll be the ones that used this moment to rebuild how work actually gets done.