Sources: Google is developing a conversational AI product, known internally as Tivoli, which aims to teach foreign languages through Google Search
Google CEO Sundar Pichai last month previewed an artificial intelligence model that he said would enable people to have open-ended conversations with technology.
Context & Ripple Effects
Weeks after announcing LaMDA as a breakthrough in natural, open-ended dialogue — the model Pichai used to frame the announcement — Google was already prototyping its first mass-market application: Tivoli, a conversational tutor for foreign languages living inside Search rather than a standalone app.
The placement matters because Search is Google's widest distribution surface, and the corpus shows the thread kept running: by 2024 Google was testing Speaking practice in Search Labs, and by 2025–26 its voice Search Live feature had rolled out broadly and then expanded to every language and region where AI Mode operates (Search Live expansion).
First-order effects
- Tivoli gives LaMDA a concrete consumer job within months of its debut — turning a research showcase into a free conversational language tutor embedded in Search.
- Users' interaction with Search shifts from discrete queries toward repeat conversational sessions, changing what engagement means for Google's core product.
Second-order effects
- Consumer apps whose core loop is chat-based language practice now compete against a zero-priced substitute distributed through the world's default search engine.
- The later Speaking practice experiment indicates Tivoli-style tutoring became a standing Search feature track, pulling Google's dialogue models into continuous real-world training through learner conversations.
Third-order effects
- If the pattern holds through the Search Live rollout and global expansion, Search evolves from a query-and-results box into a persistent voice interface where answers, tasks, and education share one conversational channel.
- Vertical single-purpose tools — language tutors among them — face structural pressure to be absorbed into horizontal assistants, mirroring how Google has folded successive capabilities directly into Search.
The trend: Google is converting Search from a lookup tool into an open-ended conversational assistant, with language education serving as one of its first mass-market entry points.