After years on Mode Analytics, our team at Prezi hit the limits of a tool that had stopped evolving: a high barrier to entry, few content creators outside the data team, and a smaller team to keep it all running. At the same time, AI had finally become mature enough to take over real analytics work.

This case study walks through how we moved to Hex, from the decision to the migration and what came after. I will cover how we decided when to switch and why one reason was never enough, how we ran a four-week proof of concept with two vendors (Omni and Hex), and what surprised even the AI skeptics on our team. Then the migration itself: triaging what to move, three waves in five months, building AI context alongside every migrated area, and running old and new dashboards in parallel to keep trust. We also look at where Hex sits in our data stack, from S3, Dagster and Amazon Athena to Hex’s MCP and its different front ends.

You will leave with a practical playbook for evaluating and migrating to an AI-Native analytics platform, and with honest lessons on what changes when AI becomes part of the stack. In our experience, the focus shifts from building dashboards and perfecting the data model to context management, basic data quality and asking good questions.