in yearly time savings from improved data access
from the time saved using Dot for analytics
that Dot is answering questions about
Duolingo is the world’s most-used language-learning platform, known for rapid experimentation and data-informed product decisions.
Duolingo wanted every employee (“Duo”) to make decisions with data—without forcing them to learn SQL, know which of 5,000+ tables to trust, or depend on busy Data Science and Data Platform teams. Non-SQL users routinely queued requests for complex analyses, creating bottlenecks and elongating the path from product question to answer. The risk wasn’t just speed; it was quality. With many similarly named tables and evolving business definitions, well-intentioned self-serve could produce inconsistent results. Duolingo needed a front door to its warehouse that was both radically accessible and uncompromising on governance.
Dot provides a text-based entry point to data directly in Slack, where Duos already collaborate. Users ask questions in natural language; Dot maps business terms to canonical tables and metrics, generates reviewable SQL, and—crucially—asks clarifying questions rather than guessing when a prompt is ambiguous. This reduces schema hunting and prevents misuse of look-alike tables.
Under the hood, Dot encodes Duolingo’s data knowledge: approved sources, metric definitions, and join logic. It enforces permissioning and routes queries to the right warehouse assets, so non-experts don’t have to memorize the data catalog. Results arrive in-thread, making it easy to validate queries, share outputs, and iterate. By shifting routine data pulls and first-pass investigations to Dot, analysts reclaim time for deeper modeling and experimentation while maintaining a single source of truth.
Dot has unlocked self-serve analysis at scale—saving 12,000+ hours, delivering an estimated 18× ROI, and cutting time-to-insight by 99%. With Slack-native, governed access to 5,000+ tables, Duos move from questions to defensible answers in minutes, while expert teams focus on higher-leverage work.
Built for data teams, helpful for all users.
Get started in minutes.