Design the Tool, Not Just the Case: An AI Automation Framework for T&S Ops
Most of us in Trust & Safety got good at one thing: finding problems. Oliver did something different. He went and designed the system that removes one.
As a T&S Investigator at Turo, Oliver-Pascal McMayoni built an AI-assisted workflow that prepares law enforcement response packages: the kind of slow, high-stakes, repetitive work that used to eat entire afternoons. It passed a full cybersecurity review. It transmits zero PII externally. And federal agencies have told Turo his packages are faster and more accurate than any other company they work with. It now runs across five countries.
Last time Oliver joined us, we talked about the journey, the fear of building something as a non-engineer. This session goes under the hood. We're pulling apart the framework and design itself: how he broke the workflow into stages, where the AI sits (and where he deliberately kept humans in the loop), the guardrails that made it safe enough to ship to law enforcement, and the reusable skeleton you could apply to your own workflow.
It's a working conversation. You'll leave with a way of thinking, not just a story.
What we'll get into:
The actual architecture — intake → process → package → output — and how to spot which of your workflows is worth automating
Where AI belongs in the pipeline, and the human-in-the-loop boundaries that matter
How "zero PII transmitted" and a passed cybersecurity review get designed in, not bolted on
What changed in the design once it rolled out to other people across five countries
The reusable framework
Built to be interactive:
Two open-floor Q&A windows (not just questions saved for the end)
A guided breakout where you'll map one of your own workflows into Oliver's framework and find the single stage worth automating
Who this is for: Trust & Safety investigators, ops practitioners, and anyone who's ever closed their laptop thinking "there has to be a better way to do this."
Come with one repetitive workflow you own. Leave seeing it as a system you could redesign.
Space is intentionally kept small so the room stays a conversation. Register to save your spot