Ask questions. Get charts. No SQL required.
Natural language to insight in seconds.
Claude
GPT-4o
Snowflake
BigQueryAsk a question
Show me revenue by region, last quarter vs this quarter
SQL generated · chart ready
Text-to-SQL via semantic layer
The AI generates SQL against your defined metrics and dimensions — so "revenue" always means the same thing, no matter who asks.
- SQL generated from semantic layer, not raw schema
- Metric logic and business rules baked in
- Works with Snowflake, BigQuery, Redshift, Postgres, and 47 more
Safe execution via MCP
Every AI-generated query runs through MCP with scoped warehouse credentials — the AI model never sees a raw connection string.
- Warehouse credentials stay in the D23 vault, never passed to the AI
- MCP enforces row-level security on every AI query
- Full audit log of every AI-generated SQL execution
Claude + OpenAI, your choice
D23 supports both Claude (Anthropic) and OpenAI models. Switch models per workspace or let D23 route to the best one for each query type.
- Claude and OpenAI GPT-4o supported today
- Model routing based on query complexity
- Bring your own API key for cost control
Works with your warehouse and your cloud
How ai analytics works
- 1
Ask in plain English
Type "Show me revenue by region last quarter vs this quarter" in the D23 AI chat or Superset SQL Lab.
- 2
AI writes the SQL
Claude or OpenAI generates warehouse-specific SQL using your semantic layer definitions — correct table names, metric logic, date filters.
- 3
MCP runs it safely
The query executes via MCP with scoped credentials. Raw warehouse passwords never touch the AI model.
- 4
You get a chart
D23 picks the right chart type automatically and renders it. You can edit the SQL, save the chart, or share the dashboard.
See it in context
The ad-hoc request
Before
A product manager emails the data team: "Can I get churn by cohort for the last 6 months?" They wait two days for a CSV.
With D23
The PM opens D23 AI chat, types the question, and gets a retention chart in 30 seconds — filtered to their product area by RLS.
Anomaly in the dashboard
Before
A spike in the revenue chart triggers a Slack message: "What happened Wednesday?" Three people dig through logs for an hour.
With D23
D23 AI detects the anomaly automatically, surfaces a natural-language explanation, and links to the relevant breakdown chart.
What teams use it for
Self-serve for non-technical teams
Finance, marketing, and ops ask data questions in plain English and get charts without writing a line of SQL.
AI-generated dashboards
Describe the dashboard you want and D23 AI builds a first draft — charts, layout, and filters — for your analyst to refine.
Anomaly detection & alerts
D23 monitors key metrics and sends a natural-language alert when something looks wrong, with context and a suggested query.
The numbers behind modern BI
You don't have to take our word for it. Here's what analysts and data teams are reporting right now.
73%
of organizations say data accessibility is the top barrier to becoming data-driven, per Forrester.
5×
faster time-to-insight for teams using self-serve BI versus those relying on a central data team for every query.
60%
of data and analytics leaders say their BI tools are too complex for business users without SQL skills.
Most teams cut their BI cost by half and ship dashboards in days, not quarters.
D23 estimate, based on typical data-team workloads: hours saved on Superset ops, faster dashboard delivery, and self-serve analytics that no longer requires a dedicated analyst for every question. Your mileage depends on warehouse size and how many teams need access to data.
It already works at this scale
Not a demo. A team in the same kind of work, with results they published.
Nielsen
Financial Services · analytics
Nielsen migrated its enterprise BI stack to Apache Superset, reducing per-seat BI licensing costs by more than 60% and enabling its data teams to ship new dashboards in days instead of weeks.
60%+
BI licensing cost reduction
days
to ship a new dashboard
1 platform
replacing multiple BI tools
The strongest results come from teams with a connected warehouse, defined metrics, and dashboards their stakeholders actually use. That is exactly what D23 delivers.
30 sec
from question to chart with AI text-to-SQL
Every team member gets answers in seconds — your analysts focus on the hard questions.
Explore more of D23
Managed Apache Superset
Zero-downtime upgrades. No infra tickets. Just dashboards.
Embedded analytics
Your brand. Your data. Invisible infrastructure.
Self-serve BI
No SQL. No tickets. No waiting.
Security & governance
SSO, row-level security, and audit logs — ready on day one.
Semantic layer
One source of truth for every number in the business.
Turn every team into a data team.
Get managed Apache Superset, the dashboards your business needs, and AI on top, without hiring a data team.