Our Blog

Looker Explore Assistant Upgrades: June 2024 Edition

The Looker Explore Assistant has some exciting new features released this week!

If you haven't heard of it, this Looker extension integrates LLMs (large language models) hosted on Vertex AI, creating a natural language experience powered by Looker's modeling layer.

In simpler terms, it allows you to generate a Looker Explore query using natural language and see the results visualized. Want to learn more? Check out the open source code here.

If you're already testing or using the Explore Assistant, keep reading to discover the latest updates!

1. Multi-Turn Chat: Engage in a Conversation with Your Data

Engage in a natural conversation with your data using the Explore Assistant's enhanced chat feature. Ask questions, seek clarification, or even request a summary of your findings—all within the chat window. This dynamic interaction streamlines collaboration between developers and stakeholders, empowering them to fine-tune queries independently and effortlessly.

The extension app further enhances your workflow by providing a set of pre-loaded sample questions right on the main page. These serve as a launchpad for your queries, demonstrating various ways to interact with your data and reducing the time spent crafting initial inquiries.

2. A More Secure Explore Assistant

Version 2.1 bolsters your data security with a new shared secret, providing an additional layer of protection for the Explore Assistant and its extension. We've also implemented Terraform to manage tokens more securely within GCP Secrets Manager, further safeguarding your sensitive information.

3. A Major Overhaul for Power and Efficiency

Version 2.0 is a groundbreaking release for the Explore Assistant, bringing a host of under-the-hood improvements that enhance performance, usability, and maintainability.

Key Technical Updates:

• Extension Examples: Introduces a script that loads examples directly into the database, simplifying setup and maintenance, and making it easier for users to get started and discover new ways to query their data.

• Styling Upgrades: Migrates to Sass for more maintainable and structured stylesheets, enabling easier customization and a more polished look.

• Backend Consolidation: Centralized Terraform scripts for better organization and consistency, making it easier to manage infrastructure and deploy updates.

• Cloud Function Interface: Standardized the Cloud Function interface to match ML.GENERATE_TEXT, simplifying usage and integration with other tools.

• Modular Backend: Separates backend modules for improved maintainability and flexibility, allowing for easier updates and customization of specific components.

• State Management: Adopts Redux Toolkit for centralized and predictable state management, leading to more reliable application behavior and easier debugging.

• Backend Selection: Enables environment-variable-based backend selection for easier switching between development and production environments, streamlining the development workflow and deployment process.

These updates are just the beginning of Google's and Data Driven’s commitment to making data exploration more accessible and intuitive for everyone. By bridging the gap between natural language and complex data queries, the Explore Assistant is at the forefront of the conversational BI paradigm.