Webinar Recap: From Bottleneck to Infrastructure

Most legal teams didn't build their marketing review process for the world they're operating in now. They built it for a handful of campaigns a year—reviewed manually, through a single channel. Then the content tsunami arrived. Google reported that advertisers used Gemini to generate nearly 70 million creative assets in late 2025 alone—a 3x year-over-year increase. The process hasn't kept pace.

In our recent webinar, From Bottleneck to Infrastructure: How Legal Teams Can Build a Marketing Review Function That Scales, we brought together legal operations leaders and General Counsels to talk honestly about what's breaking and where to start. Panelists included Sheila Dusseau, Head of Global CLO Operations and Innovation at Ferring Pharmaceuticals; Antonia Walters, Head of Legal Engineering and Operations at Blee; and a VP of Legal from Mastercard—all moderated by legal industry analyst Ari Kaplan.

While you may have missed the live discussion, we captured the most important insights below. If you'd like to watch the full session, you can access the recording here.

Visibility Disappears Without a Single Path

The volume problem is real. But according to Sheila Dusseau, volume is only part of the story. The deeper issue is fragmentation. Content is being created everywhere, marketing teams, agencies, influencers, product teams, with no single intake channel for legal to track.

"When you can't track the data, when you don't know what's in flight, you don't know who's owning it, you don't even know what risk tier it's in," Dusseau said. "Then you don't have a scalable function."

The result is a compliance function that's reactive by design. Without visibility into what's moving through informal channels, legal teams are always catching up and the gap between content volume and review capacity continues to widen.

Fix the Process Before You Touch the Technology

When something slips through legal oversight, the instinct is to blame the tool or add a reviewer. Antonia Walters argues the problem is almost always upstream. "The guidance might exist. The playbook might exist. But it might not live in the place where marketers work, where they spend their day to day."

The point isn't that people are trying to circumvent legal. It's that the process wasn't designed around the real content lifecycle. The stakes are real: FTC civil penalties run up to $53,088 per violation—and that's per asset. A campaign with hundreds of non-compliant influencer posts can quickly become an existential financial risk. And if the process is broken, technology won't save you. As Dusseau put it: "AI isn't gonna fix a broken process. It's just gonna make it a faster broken process."

Both Walters and Dusseau were clear on sequencing: understand your workflows, define your sources of truth, and establish where human judgment is required before introducing any AI layer. "Data is your foundational enabler," Dusseau said. "You've got to get your data right and in good form." Getting your data in order isn't optional. It's the foundation everything else runs on.

Where AI Genuinely Helps and Where It Doesn't

Once the process is solid, AI can do meaningful work. Walters described three areas where it adds real value: visibility into content that never reached the formal review queue; consistency across reviewers applying the same risk standards; and triage—surfacing risky claims, flagging missing disclosures, comparing drafts against approved language.

What AI can't do is make business risk decisions. It can tell you a claim is risky. It can't tell you whether the business should take that risk. "It's not a fail safe," Walters said. "And it's not an answer to everything."

For legal teams evaluating AI tools, Walters recommends treating them less like standalone software and more like operating infrastructure. The right questions aren't about what the tool flags, they're about how it connects to your existing sources of truth, how it handles uncertainty, who controls the rules, and what implementation actually requires.

The Biggest Risk Isn't Hallucinations, It’s Unmanaged Adoption

Dusseau closed with a challenge to one of the most common assumptions about AI risk. "It's not about the hallucinations. We know that's a problem, but the bigger risk—it's actually unmanaged adoption. Without guidance and enablement, people are gonna use unapproved tools in unapproved ways."

The teams getting this right aren't necessarily the ones with the most sophisticated tools. They're the ones who brought people along, building AI literacy before deploying the technology, starting with practical use cases, and treating training as an ongoing operating model rather than a one-time event.

As Walters put it, if the tool is only designed for legal, marketers will avoid it. If it's only designed for speed, legal won't trust it. The goal is infrastructure that works for both. As Dusseau framed the shift: "Legal ops is shifting away from support and more into legal engineers, system leadership, designing experiences."

The function is changing. The tools need to change with it.

Watch the full session to hear what Walters thinks a scalable marketing review function looks like three years from now.