Skip to main content
The second step in Adapt’s Ask-Understand-Act framework. After you ask a question, Adapt investigates across your connected data sources to provide trusted, cited answers.

​
How Adapt Investigates

When you ask a question, Adapt’s AI agent:
1

Plans the approach

Determines which data sources to query and what information is needed
2

Gathers data

Queries your connected tools—Snowflake, HubSpot, Stripe, etc.
3

Analyzes results

Processes the data, performs calculations, and identifies patterns
4

Synthesizes answer

Combines findings into a clear, comprehensive response with citations

​
Trusted Answers

​
Source Citations

Every answer includes citations to the data sources used: 
Your MRR grew 12% last month, from $98,400 to $110,200.

Sources:
• Stripe subscription data (as of today)
• 47 active subscriptions analyzed
You can always trace an insight back to its source.

​
Data Transparency

When Adapt queries your data, you can see:
  • Which tools were accessed
  • What queries were run
  • Raw data that informed the answer
Click “Show details” on any response to see the full investigation.

​
Multi-Source Analysis

Adapt can combine data from multiple sources in a single answer: Question: 
Which customers are at risk of churning?
Adapt investigates:
SourceData Retrieved
StripePayment history, failed charges, plan changes
HubSpotLast contact, support tickets, engagement score
Product DBLogin frequency, feature usage, activity trends
IntercomRecent conversations, sentiment
Combined insight: 
5 accounts show churn risk signals:

1. Acme Corp - No login in 45 days, 3 open support tickets
2. Beta Inc - Downgraded plan, payment failed twice
3. Gamma LLC - Usage dropped 80% this month
...

​
Analysis Capabilities

​
Data Processing

Adapt can process and transform data:
  • Aggregations (sum, average, count)
  • Grouping and segmentation
  • Time-series analysis
  • Percentage calculations
  • Trend identification

​
Statistical Analysis

For deeper questions, Adapt performs statistical analysis:
  • Correlation analysis
  • Cohort comparisons
  • Significance testing
  • Forecasting and projections

​
Code Execution

When complex analysis is needed, Adapt runs code in a secure sandbox: 
# Adapt might run analysis like this:
import pandas as pd

# Load and process data
df = pd.read_csv('sales_data.csv')
monthly = df.groupby('month')['revenue'].sum()
growth = monthly.pct_change()

# Calculate insights
avg_growth = growth.mean()
best_month = growth.idxmax()
You can see the code that was executed and verify the logic.

​
Visualization

Adapt automatically generates visualizations when helpful:

​
Charts

Request specific chart types: 
Show me a line chart of revenue by month

Create a bar chart comparing regions

Visualize the distribution of deal sizes

​
Tables

Complex data is formatted into sortable, searchable tables:
CustomerRevenueGrowthRisk Score
Acme Corp$45,000+12%Low
Beta Inc$32,000-5%Medium
Gamma LLC$28,000-15%High

​
Dashboards

For comprehensive views, Adapt can create multi-visualization dashboards combining charts, tables, and key metrics.

​
Web Research

For questions requiring external information, Adapt searches the web: 
What are the latest trends in B2B SaaS pricing?
Adapt will:
  1. Search for recent articles and reports
  2. Extract relevant information
  3. Summarize findings with source links
  4. Combine with your internal data if relevant

​
Deep Research

For complex research questions, Adapt can spawn sub-agents to investigate in parallel: Question: 
Do a competitive analysis of our top 3 competitors
Adapt’s approach:
  • Sub-agent 1: Research Competitor A
  • Sub-agent 2: Research Competitor B
  • Sub-agent 3: Research Competitor C
  • Main agent: Synthesize findings into comparison
This parallel approach delivers comprehensive research faster.

​
Knowledge Base

Adapt maintains a knowledge base of your organization’s context:
  • Definitions: How you define metrics like “active user” or “MRR”
  • Business rules: Your fiscal year, quota structures, etc.
  • Preferences: How you like reports formatted
  • History: Learnings from past conversations
This context helps Adapt give more relevant, accurate answers.

​
Verification

For critical insights, Adapt can verify claims:
  1. Cross-reference multiple data sources
  2. Check for data quality issues
  3. Flag any inconsistencies
  4. Note confidence levels
Revenue grew 15% last quarter.

Verification:
✓ Confirmed in Stripe data
✓ Matches Snowflake records
✓ Consistent with HubSpot deal data
Confidence: High

​
Next Steps

Once Adapt understands your question and provides an answer, you can:

Ask Follow-Ups

Drill deeper with follow-up questions—context is maintained

Take Action

Have Adapt act on the insights—draft emails, update records, etc.