Use caseData Analysis Agent

Your data team, on demand in Slack

The exec asks 'what's our paid CAC by channel this quarter?' — Dench writes the SQL, runs it on your warehouse, charts the answer, and posts it. Without joining a Slack thread.

Start freeSee the workspace
#exec-questions/Data Analysis Agent agentlive
DenchAgent10:34

Re: 'What's our Q3 paid CAC by channel?'

Pulled from dbt model fct_paid_acquisition, joined to closed-won opps. Window: Jul 1 – Sep 30, blended attribution. Click the chart for the SQL.

TikTok58
Google87
Meta124
LinkedIn342
Promote to dashboardOpen SQL
Answered from dbt model · 1.4s query timeDench
Built for

Data teams drowning in 'can you pull a quick number'

  • Slack DMs from execs pile up faster than you can run SQL
  • Dashboards nobody opens, ad-hoc nobody scopes
  • Exec team is just using ChatGPT for 'analysis' anyway

What the Data Analysis Agent agent actually ships.

  • Self-serve SQL for everyone

    Writes, runs, and explains queries against your warehouse. Stakeholders ask in English; Dench replies with the chart.

  • Trusted by the data team

    Uses your dbt models, semantic layer, and metric definitions — answers match what BI would have run.

  • Dashboards that update themselves

    Auto-refreshes the dashboards execs actually open. Annotates anomalies in plain English.

How it works

When the CEO asks 'what's our CAC by channel?'

  1. Step 01

    Question lands in Slack

    Plain English. No JIRA ticket, no spec doc, no 'let me get back to you'.

  2. Step 02

    Dench writes the SQL

    From your dbt models and metric definitions — not a hallucinated table name.

  3. Step 03

    Result charts itself

    Posted back to Slack with the chart, the SQL, and the caveat about the data window.

  4. Step 04

    Promotes to a dashboard

    One click moves it to Mode, Hex, or Metabase. Now everyone has it.

What it ships

What the Data Analysis agent actually ships.

  • Answer ad-hoc data questions in Slack

    SQL written, run, charted, posted

  • Read your dbt models and metric layer

    Answers match the source of truth

  • Detect dashboard staleness and fix it

    Refresh pipelines, flag broken charts

  • Generate weekly KPI reports per team

    Posted to the team's Slack channel

  • Catch anomalies in the metric layer

    Slack alert when CAC, churn, or DAU drift

  • Suggest dbt model improvements

    From the queries it sees stakeholders ask

  • Build the executive monthly business review

    From the warehouse, ready for the doc

  • Convert ad-hoc queries to reusable models

    When the same question lands three times

  • Document tables and columns from real usage

    Auto-generated dbt docs, kept current

  • Surface unused dashboards

    Stop maintaining what nobody opens

Integrations

Built for the modern data stack.

  • Snowflake
  • BigQuery
  • Redshift
  • dbt
  • Mode
  • Hex
  • Metabase
  • Looker
  • Slack
  • Linear

What Garry Tan, CEO of Y Combinator, says about Dench

Placing agent power on your own computer empowers every user and I’m so here for that. dench.com/claw
Garry Tan

Garry Tan

CEO of Y Combinator

600K+ followers

Questions you'd actually ask

Before you switch.

Dench is grounded in your dbt models, metric layer, and warehouse schema. It only writes SQL against real tables, with real joins, using your defined metrics. Wrong answers fail loudly, not silently.

Reclaim Friday from the Slack DM pile.

Connect Dench to your warehouse and dbt. Watch the first 'quick question' answer itself.

Try Dench