Indent is your new AI data analyst
Get instant, accurate answers to any data question using AI that understands your data warehouse, BI tools, and code.
Connect to any data store
BigQuery
ClickHouse
Databricks
MySQL
PostgreSQL
Answers without waiting for the data team
Ask a question in plain English and Indent goes to work—giving you the answer right away.
A
How many users signed up in Japan from our Facebook campaign last week, and how were the signups distributed throughout the week?
Indent handles complex questions. Just like a good colleague, Indent will clarify your intent and break large questions into manageable pieces.
Analyzing...
Referenced 12 observations related to users and marketing
Read:
users
Read:
analytics.marketing
I'll help you analyze the signup data from last week's Facebook campaign in Japan. Let me query the users table and examine the daily distribution.
Query:
japan-facebook-signups
Query:
japan-facebook-signups
Reasoning...
Indent shows every step of its work. See exactly how Indent got to its answer, so you can audit and trust the final output.
342 users signed up last week in Japan from the Facebook campaign.
The signups were distributed throughout the week, with the earliest on October 6th and the latest on October 12th. The data is current through October 15th, so this count is complete for the requested period.
View explanation
I queried the users table filtering by created_at between October 6th, 2025 and October 12th, 2025. I filtered for users from Japan (country = 'JP') and acquisition source as Facebook (utm_source = 'facebook'). Following standard practice for user metrics, I excluded bot users (bot_user = false OR NULL) and the Indent internal organization (organization_id != 3) as specified in the observations.
I verified data recency with a second query which confirmed that user signup data is current through October 15th, 2025, ensuring the last week count is complete and accurate.
Indent's answers are thorough. Indent uses multiple queries, visualizations, and explanations to give you a complete picture.
Integrates with dashboards and codebases
Indent understands your entire data stack. Context from dashboards and codebases allow Indent to improve answer quality beyond what's possible with tables alone.
While working on a question, Indent references your dbt lineage, checks your Airflow logs, and validates against your BI definitions . This context is the difference between a guess and an answer.
Indent never forgets a thing.
Just like onboarding a new colleague, you'll teach Indent a thing or two about what makes your data unique. The difference is, you only ever need to tell Indent once.
2 new observations
Review now
Analysis complete!
Key findings for users signed up in the last 2 months:
Conversion Rates:
• 30.8% of users created their first chat (28 out of 91 users)
• 24.2% of users reached code generation (22 out of 91 users)
Time to First Chat:
• Average: 78.2 hours (3.3 days)
• Median: 1.6 hours (much faster than the average, some users start immediately while others take much longer)
Time to First Code Generation:
• Average: 108.7 hours (4.5 days)
• Median: 21.6 hours (0.9 days)
The visualizations show the distribution patterns and user journey stages. Most users (69%) haven't created any chats yet, while about 7% created chats but haven't generated code. The timing distributions are right-skewed, with medians much lower than averages, indicating many users engage quickly while others take significantly longer.
Indent is constantly improving itself. Every time Indent comes across new information, it proposes updates to its knowledge base for you to review.
"Premium signups"
users.tier = 'premium'
"Conversion rates"
SUM(paid_users) / SUM(trial_users)
Indent learns to speak like you. Indent will adapt to concepts like the steps of your onboarding funnel so you can talk to it as naturally as a teammate.
Analysis Complete!
Key Findings for users signed up in the last 2 months:
Conversion Rates:
• 30.8% of users created their first chat (28 out of 91 users)
• 24.2% of users reached code generation (22 out of 91 users)
Time to first chat:
• Average: 78.2 hours (3.3 days)
• Median: 1.6 hours (much faster than avg, some users start immediately while others take much longer)
Time to first report generation:
• Average: 108.7 hours (4.5 days)
• Median: 21.6 hours (0.9 days)
The visualizations show the distribution patterns and user journey stages. Most users (69%) haven't created any chats yet, while about 7% created chats but haven't generated a report. The timing distributions are right-skewed, with medians much lower than averages, indicating many users engage quickly while others take significantly longer.
Approve
Reject
Analyst
Indent gives you tools for corrections. Built-in review flows let your team flag inconsistencies and let the data team double-check every query and result.
Stop waiting for data answers. Start getting them instantly.
Talk with us
Company
LInks