Dust's latest capability, structured data analysis with table queries, marks a significant leap forward to help anyone in the company make the most of data. This feature enables AI assistants to generate and execute SQL queries on structured data sources like CSV files, Notion databases, and Google Sheets, opening up new quantitative analysis avenues.

You were explicit in your fondness for Dust's swift semantic search capabilities for your company's knowledge base, and also in your wish for a powerful structured data analysis tool. Dust do it. Dust now introduces SQL query generation for structured data analysis, distinct from, but complementary to, our semantic search functionality.

Using Table Queries

Dust offers two primary methods to leverage table queries:

1. Manual Table Creation: Users can upload CSV files or utilize Dust’s API to create tables within folders.
2. Automatic Table Discovery: If you created a Connection with Google Drive or Notion, Dust seamlessly discovers tables from Google Sheets and Notion Databases. This integration ensures that your data is always up-to-date and ready for analysis without the need for manual intervention.

What are Real-Life Applications of Structured Data in Dust

Structured data analysis with Dust can transform workflows and power new use cases across teams:

Pipeline Management in a Notion DB:

Track your interactions with candidates, freelancers, and partners by querying your Notion database for real-time insights.

Meet John from the Marketing team.

John needs to track their engagement with freelancers across various projects. John created a custom assistant @great_freelancers.

@great_freelancers is powered by GPT4 Turbo and has access to the team's Marketing Notion Database, which records all the ongoing relationships with freelancers working with the team on corporate comms, public affairs, social media, events, campaigns, PR, and videos.

John can instantly answer queries like "How many freelancers are we discussing with for project X?" or "Who it our point of contact for X".

Project Repo in Google Sheets

Manage your projects efficiently by asking your Google Sheet detailed questions regarding project statuses and metrics.

Meet Stan, and has access to the Project Prioritiation Google Sheet, which records all the backlog and ongoing projects with associated DRIs and project sizes who managed the Eng team.

Stan craves for speed and wants to find actionable solution to ensure the Eng team velocity. Prioritization is a morning routine. Stan created @projects.

@projects is powered by GPT4 turbo and has access to the Project Prioritisation Google Sheet, which records all the backlog and ongoing projects with associated DRIs and project sizes.

Stan can effortlessly answer questions regarding project status, resource allocation, or deadlines with precision and speed.

ARR Analysis

Update your investors or company memos with the latest ARR figures by querying your ARR data stored in CSVs or Google Sheets.

Meet Casey from the Finance Team

Casey wants to keep investors and the leadership team informed and aligned with the company’s financial status. To streamline this process, Casey created a custom assistant, @ARR_insights.

@ARR_insights is powered by GPT-4 Turbo and has access to the company's financial performance data stored in both CSV files and Google Sheets.

With @ARR_insights, Casey can swiftly answer queries such as "What was the total ARR at the end of Q2?" or "How does the ARR of Product A compare to Product B in the first half of the year?" .

Conclusion:

Dust's introduction of structured data analysis through table queries is a game-changer for smart companies willing to leverage AI for both semantic search and quantitative analysis.

We're committed to enabling every team within your organization to harness the most advanced LLMs, crafting a personalized AI ecosystem.

If you are a team ready to harness the power of structured data, sign up for our Pro Plan today. Enterprises with more than 100 employees are invited to reach out directly at team@dust.tt. Join our AI builder community, and let's make work, work better.