Gina Jenks

Self-service has been promised before. Every BI vendor put it on their front page a decade ago, most data leaders tried it, and the ticket queue never went away. The skepticism is earned. The usual diagnosis is that the tools weren’t good enough. But something subtler is going on. There’s always a gap between what someone asks for and what they actually need. By the time a request lands in your queue, it’s already been compressed: a problem someone lives with daily, translated into a shape they think a data team can service. Good analysts close that gap. They just have to do it one ticket at a time. Synthesia’s Gina Jenks joins Omni co-founder Jamie Davidson to talk about what happened when Synthesia stopped closing that gap ticket by ticket. What they built, which of the usual blockers turned out not to matter, and what people made once they didn’t have to ask first.

What happens when you let your entire 700+ employee organisation loose on AI tooling using your data with one month’s notice? At Synthesia, the data team are driven by a mission to provide self-serve data for any purpose, not just dashboards and reporting. With the introduction of AI tooling, non-technical users were no longer dependent on others for their data. The biggest surprise came from what people built: not tools to inform decisions, but tools enriched with data to run their day-to-day operational work. In this session, Gina Jenks, AI Knowledge & Analytics Manager at Synthesia, discusses what happens when a data team gets out of the way of its users, trusts stakeholders to self-correct, and designs for high levels of ambiguity rather than against it. Drawing on real examples from across the business, Gina explores why operational use cases raise the bar for freshness and reliability, why trust beats control when you can no longer check every answer, and how Synthesia uses Fivetran and dbt Labs to build the trusted data foundation that supports AI-powered self-service at scale. Attendees will walk away with a clearer view of what AI-ready self-serve actually looks like in practice, and where to start.