Migrations need to be fast and seamless. But at scale, they’re often anything but – especially when many teams are involved and coordination overhead balloons. What if you could make migrations self-serve, so they just happen, without endless follow-ups or hand-holding?
At Airbnb, we’ve been working AI into our migration toolkit to rethink migrations from the ground up, guided by two tenets: “Complete to 100%” and “Be a host” – taking responsibility for migrations on behalf of teams.
In this talk, I’ll share how we built tooling, infrastructure, and guidance to enable rapid, large-scale AI-driven migrations. Through some case studies, you’ll learn how these principles are letting us scale migrations across the company, the new challenges we faced (like review bottlenecks), and practical lessons for anyone looking to accelerate migrations in their own organization.
Key takeaways:
- How migration tenets like “be a host” and “complete to 100%” are foundational for accelerating migrations with AI by changing the planning, ownership, and tooling approach
- The specific tools and infrastructure we’ve built to scale AI-accelerated migrations and make them self-serve and repeable
- The key points we use to train our migration owners to help deliver successful AI-accelerated migrations
- And some of the bottlenecks and hurdles we’ve hit doing this (some we’ve addressed, some we haven’t yet)
