No warehouse required

The AI data analyst that shows its work 

Ask questions across the databases you already run. Every number comes back with its rows and reasoning, and your team's definition wherever one applies.

Churn analysis

Why did churn spike in March?

  1. Reading billing_prod.subscriptions
  2. Applying definition: churned account
  3. Comparing with the 12-month baseline

Cancelled accounts per month

    Connects to the databases and apps you already run

    • PostgreSQL
    • Snowflake
    • MySQL
    • SQL Server
    • BigQuery
    • MongoDB
    • Oracle
    • Databricks
    • ClickHouse
    • Salesforce
    • MariaDB
    • Supabase
    • SAP
    • Neon
    • CockroachDB
    • TimescaleDB

    Watch a question become an answer

    1. Connect

      Point Unveilix at the databases you already run. Read-only, nothing copied.

    2. Ask

    3. Trace

    4. Answer

    Postgresbilling_prod
    Snowflakeanalytics
    MongoDBproduct_events
    Why did churn spike in March?
    Halden FreightStarter
    Northpine DentalStarter
    Corvo StudioGrowth
    Bramble & LoweStarter
    Linden LabsStarter
    Ostara HealthGrowth
    March: 96
    Definition96 rowsInterpretation

    Pin any answer to a dashboard that stays current

    Every tile is a saved question. It reruns against your live databases and keeps its citations.

    Revenue overview

    Sample data, refreshed hourly

    Net revenue, last 12 months

    $5.18M

    39.0%monthly run rate vs Oct

    Active customers

    18,406

    21.0%vs Oct

    Month 3 retention

    84%

    Average of Apr to Jun cohorts

    Net revenue by region

    North AmericaEuropeAsia Pacific
    Monthly net revenue by region, last 12 months, in thousands of dollars
    MonthNorth AmericaEuropeAsia Pacific
    Oct$182K$121K$64K
    Nov$188K$119K$67K
    Dec$195K$126K$71K
    Jan$191K$130K$70K
    Feb$204K$128K$76K
    Mar$210K$135K$81K
    Apr$216K$139K$85K
    May$223K$137K$90K
    Jun$219K$144K$94K
    Jul$231K$149K$97K
    Aug$238K$151K$103K
    Sep$246K$156K$108K

    Revenue by channel

    • Direct sales$2.07M 40%
    • Partners$1.30M 25%
    • Marketplace$933K 18%
    • Self-serve$881K 17%

    Customer retention by signup cohort

    Hover a cell for its value
    Share of each monthly signup cohort still active, months 1 to 5
    CohortM1M2M3M4M5
    Apr91%86%83%81%80%
    May92%87%85%83%
    Jun90%86%84%
    Jul93%89%
    Aug94%

    Check the number before you act on it

    A governed answer arrives with three things attached: the approved definition, the rows and the interpretation.

    96 accounts cancelled in March, against a usual 42.

    Churned account96 rowsPaused excluded

    The definition

    Churned accountApproved
    Used in 214 answers

    Owned by Customer Success · v3, edited Feb 2

    count(distinct account_id)
where subscription.status = 'cancelled'
  and cancelled_at in reporting_month
  and not account.is_paused
    Same meaning inFinanceCustomer SuccessSales

    The rows

    96 rows from billing_prod.subscriptions

    AccountPlanCancelledMRR
    Halden FreightStarterMar 3$490
    Northpine DentalStarterMar 4$490
    Corvo StudioGrowthMar 9$1,250
    Bramble & LoweStarterMar 12$490
    and 92 more rows

    The interpretation

    Why did churn spike in March?cancelled accounts · Mar · excl. paused
    • Read churn as cancelled subscriptions, not downgrades
    • Period March 1 to 31, by cancellation date
    • Excluded paused accounts and internal test workspaces
    • Compared with the average of the other 11 months

    One analyst layer between your team and your data

    Unveilix sits on the databases you already run, applies the context your team agreed, and hands back answers anyone can check.

    How security and data access work

    Your team

    Unveilix analyst

    Business context

    Your databases

    Data flows up, answers flow back with their receipts

    • 0rows copiedQueries run where your data already lives, read-only. Nothing moves into a warehouse first.
    • 3receipts on a governed answerThe approved definition, the rows and the interpretation travel with each governed number.
    • 1definition per termDefinitions your team approves once are used on every question that needs them.

    Answers on day one, not after a data project

    No warehouse, no migration

    Query your CRM, product and finance databases together without copying a row anywhere first.

    1. 1Connect
    2. 2Read schema
    3. 3Ask
    billing_prodRead-only · TLSConnecting

    Reading schema in place

    • billing.subscriptions48,210 rows
    • billing.invoices312,904 rows
    • crm.accounts9,876 rows
    • crm.opportunities21,450 rows
    • events.sessions4.2M rows

    Tables found

    0

    Rows copied to a warehouse

    0queried in place

    Ready in 4 minutesAsk: What's our churn by plan?

    Your data stays put

    Read-only connections, a secure client for databases behind your firewall, and a log of every question.

    Read-only connection, audit log

    • Priya Nair asked about churn by planjust now
    • Mei Tanaka asked about refunds in Europe2m ago
    • Jonas Berg edited the Active customer definition5m ago

    One definition per term

    Agree once on what active customer means. Two people asking the same question get the same number.

    FinanceHow many active customers do we have?4,212
    SalesActive customers this month?4,212

    Same number · Active customer v3

    Built for the whole team

    Invite colleagues, give each one access to the right sources, and share threads and dashboards.

    MemberBillingAnalyticsProduct
    PPriya
    TTomás
    AAda

    Less time wrangling data, more time deciding

    Compare Unveilix with text-to-SQL and BI tools

    The usual route

    Build a warehouse
    Model the data
    File a request
    Wait for SQL
    Get a number
    Scoping the warehouse…

    Weeks of setup, and the person who writes SQL becomes the bottleneck.

    With Unveilix

    Connect
    Ask
    Check the citations
    Connected…

    Everyone asks directly and the answer shows its work.

    Ask your own data a question today

    Connect a database in minutes. Every answer comes back with its rows and reasoning, and your team's definition wherever one applies.