Skip to main content

​
Overview

FigAI Analyst is your AI-powered data analyst that transforms natural language questions into data insights. Ask questions in natural language and get instant answers with charts, tables, and analysis—no SQL or technical skills required. Key Capabilities:
  • Natural language questions across all connected data sources
  • Build and modify workflows
  • Dashboard summaries
  • [COMING SOON] Chart generation (bar, line, pie, table) that can be added to dashboards
  • Conversational follow-ups to refine results

​
Getting Started

1

Access FigAI

From any screen on the app, click FigAI diamond icon at the bottom right corner of the screen. Ensure you have at least one connected data source (database, connector, Virtual Table, or upload).
2

Ask your question

Type in natural language. Examples:
  • “What were our total sales last month?”
  • “Show me top 10 customers by revenue”
  • “How has revenue changed over time?”
3

Review results

FigAI returns an answer summary, visualization (when appropriate), data preview, and the SQL query used (for transparency).
4

Click prepopulated follow up questions or ask another question

Want to dig deeper? FigAI provides recommendations for follow up questions.

​
Asking Effective Questions

​
Question Structure

What you want to know + About which data + Time period/filters (optional) Good Examples:
  • “What were total sales by product category last quarter?”
  • “Show me customers who haven’t purchased in 90 days”
  • “Compare revenue this year vs last year by month”
Less Effective:
  • “Sales” (too vague)
  • “Show me everything” (too broad)
  • “Is the business doing well?” (too subjective)

​
Question Types

Aggregation: “What’s our total revenue this month?” “How many orders yesterday?” Trends: “How has revenue changed month over month?” “Show daily users for last 30 days” Rankings: “Who are top 10 customers by revenue?” “Which products have lowest sales?” Filtering: “Show orders from New York in 2024” “List transactions over $1,000” Comparisons: “Compare revenue by region” “What’s conversion rate mobile vs desktop?” Distribution: “What’s the distribution of order values?” “Show customer count by age group”

​
Refining Analysis

​
Follow-Up Questions

FigAI remembers context within each conversation. Build on previous questions:
  1. “What were total sales last month?” → “$45,230”
  2. “Break that down by product category” → Bar chart by category
  3. “Which products in Electronics sold most?” → Top electronics products
  4. “Show daily trend for those products” → Line chart

​
Modifying Questions

  • Add filters: “Show sales” → “Show sales for California customers”
  • Change time: “What’s revenue?” → “What’s revenue in Q1 2024?”
  • Adjust granularity: “Revenue over time” → “Revenue by month last year”
  • Change metric: “Customer count” → “Customer count and average order value”

​
Saving Results

Click Save as Virtual Table below results. Give it a descriptive name and use in:
  • Dashboard figs for visualization
  • Workflow Input nodes for further transformation
  • Shared reports for team access

​
Common Question Patterns

Metric + Time Period: “What was [metric] in [time period]?” Metric + Breakdown: “Show me [metric] by [dimension]” Ranking: “What are the top [N] [entities] by [metric]?” Comparison: “Compare [metric] between [segment A] and [segment B]” Filter + Metric: “What is [metric] for [filtered subset]?”

​
Example Questions

Sales: “What was total revenue last month?” “Show daily revenue for last 30 days” “Revenue by product category?” Customers: “How many new customers this month?” “Customers who haven’t purchased in 90 days” “Average purchase frequency?” Marketing: “Conversion rate for email campaign?” “Leads from paid ads last week?” “Website traffic by source?” Operations: “Support tickets created today?” “Average ticket resolution time?” “Most common issue categories?” Product: “Daily active users?” “Most used features?” “Feature adoption rate for new dashboard?”

​
FigAI vs Workflows

​
Use FigAI When:

  • Quick ad-hoc analysis and exploration
  • Answering one-off questions
  • Validating hypotheses before building workflows
  • No technical skills available
  • Sharing quick insights

​
Use Workflows Instead When:

  • Complex multi-step transformations needed
  • Reusable processes for regular analysis
  • Production dashboards with scheduled updates
  • Data cleaning and quality checks
  • Joining multiple data sources
  • Custom calculations with complex business logic
Best Practice: Start with FigAI to explore and validate, then convert useful analyses to workflows for production use.

​
Limitations

FigAI Cannot: Handle complex multi-step transformations, execute custom business logic, modify data (read-only), or run scheduled automation. Use workflows for these needs.
LLMs Can Make Mistakes: Always double check the output provided by the FigAI Analyst before making any major decisions.

​
Best Practices

New Conversation for Each New Topic: Start new conversations when pivoting to a new topic. This keeps the context from the conversation relevant, and helps prevent hallucinations.
Be Specific: Include metric, data source, and time period/filters.
Start Broad, Then Narrow: General questions first, then drill with follow-ups.
Include Time Context: Specify “last month”, “this year”, “Q1 2024” explicitly.
Check the SQL: If results seem off, view SQL to verify FigAI understood correctly.