Reflections from MRMC SF 2025: Going beyond basics in marketplace trust and safety
As Persona’s trust and safety architect, I get a front-row seat to how marketplaces are evolving — the risks they’re navigating, the systems they’re building, and the gray areas in between. Last week’s Marketplace Risk Management Conference (MRMC) in San Francisco made one thing especially clear: today’s threats demand more than just checking the right boxes.
As a longtime MRMC partner, we value how the conference challenges platforms to think more critically about trust and safety. From panels on AI-generated fraud to deep dives into pig-butchering scams and platform abuse, one thing was obvious during this year’s event: the stakes are getting higher, and solutions have to keep up.
Here are a few of my biggest takeaways from the sessions I attended and people I spoke with:
Takeaway #1: Go beyond verification and rethink your onboarding holistically
Too often, onboarding gets treated like a single moment: fill out a form, upload a document, done. But leading marketplaces approach onboarding as a foundational part of their growth and risk strategy — one that extends far beyond a single interaction.
In our session with Josh Cutler, CTO of Yardstik, and Sidra Khan, head of compliance at Taskrabbit, we dug into what a more thoughtful onboarding flow can look like. One key tactic: start with cheaper, faster checks like ID verification to filter out obvious bad actors, and reserve more resource-intensive steps like background checks for users who pass the initial screen(s).
Taskrabbit’s results speak for itself: a 20% drop in background check failures and less strain on its ops team. It’s a small shift that makes a real difference in conversion, cost, and overall experience.
We’re excited to help more marketplaces adopt this approach with our new integrated background check and IDV solution, now in early access. By bringing these checks together in a single, configurable flow, we’re making it easier than ever to build smarter, safer onboarding from day one. Stay tuned for more details!
Takeaway #2: In the age of AI, identity isn’t a given
If we had a nickel for every time someone mentioned generative AI at MRMC SF… well, let’s just say we could fund a pretty robust fraud team. From cloned voices and synthetic faces to bots trained on online forums, AI-powered fraud was top of mind across nearly every panel.
We spoke about this in our own session, Digital Identity Crisis: How to Build Trust in the Age of AI, where we opened with a simple stat: 6 in 10 Americans say they’ve interacted with AI online without realizing it. That blurring between human and machine is now part of the trust equation — and it’s not one a “verified” badge can solve on its own.
The most resilient marketplaces are starting to think more holistically: layering identity checks with device signals, behavioral patterns, and contextual data to better understand who is on the other side of the screen. They’re also strategically choosing where and when to add friction, using early signals to dynamically route users instead of sending everyone through the same gauntlet.
Takeaway #3: KYB is more important than ever
While AI-powered fraud captured much of the conversation, it wasn’t the only threat top of mind for marketplaces. KYB fraud — especially from synthetic businesses — surfaced across several sessions. One standout panel, Protecting Sellers from Fraud and Abuse on Marketplaces, hosted by our customer Grubhub, explored the unique risks restaurants face, highlighting large-scale phishing campaigns that have compromised thousands of accounts and devices globally. It wasn’t all doom and gloom, though — Grubhub also shared how its Know Your Business (KYB) protocols help the platform stay ahead of these threats.
More marketplaces are now following suit, investing in proactive KYB strategies to stop bad actors before they ever get in the door. Whether it’s detecting impersonation attempts, uncovering fraud rings, or preventing account takeovers, the goal is the same: block fraud without slowing down trusted sellers.
Takeaway #4: Collaboration isn’t optional — it’s essential
The main keynote on pig-butchering scams hit a sobering note: fraud rings are growing more sophisticated, more global, and more coordinated. If platforms try to tackle organized fraud alone, they’re going to lose ground — fast. Marketplaces, law enforcement agencies, and other governing bodies need to thoughtfully collaborate if they want to stand a chance against these increasingly sophisticated fraud networks.
But the need for collaboration doesn’t stop with cross-platform intelligence sharing. In our joint session with other MRMC board members, we also talked about internal collaboration — how trust and safety, ops, legal, and product teams can work better together, share insights, and build alignment around risk tradeoffs.
The future of marketplaces in 2025 and beyond
Fraud is evolving fast, and identity is no longer a simple binary. As the lines blur between human and AI, trust and safety work needs to become more adaptive, integrated, and team-oriented.
MRMC SF was a reminder that no one has this all figured out — but we’re all asking the right questions, and we’re building toward better answers together.
Looking to strengthen your marketplace’s onboarding, KYC, or KYS strategy? Reach out to our team.
