---
title: Bits Data Analysis
description: >-
  Ask questions about your business data in natural language. Bits Data Analysis
  writes the queries, runs them against your data warehouse, and returns answers
  without writing SQL or navigating dashboards.
breadcrumbs: Docs > Bits AI > Bits Data Analysis
---

# Bits Data Analysis

{% callout %}
# Important note for users on the following Datadog sites: app.ddog-gov.com, us2.ddog-gov.com

{% alert level="danger" %}
This product is not supported for your selected [Datadog site](https://docs.datadoghq.com/getting_started/site.md). ({% placeholder "user-datadog-site-name" /%}).
{% /alert %}

{% /callout %}

{% callout %}
# Important note for users on the following Datadog sites: app.datadoghq.com, us3.datadoghq.com, us5.datadoghq.com, app.datadoghq.eu, ap1.datadoghq.com, ap2.datadoghq.com

{% callout %}
##### Join the Preview!

Bits Data Analysis is in Preview. Click **Request Access** to join the Preview program.

[Request Access](https://www.datadoghq.com/product-preview/bits-data-analysis/)
{% /callout %}

{% /callout %}

## Overview{% #overview %}

Bits Data Analysis is an agentic AI tool that lets you explore and analyze your business data through natural language. Ask questions about your revenue, customers, support, or sales pipeline. Bits Data Analysis identifies the relevant tables, writes the SQL, runs it against your connected data warehouse, and returns the results. Combine business data with your Datadog observability data in dashboards, notebooks, and [DDSQL](https://docs.datadoghq.com/ddsql_editor.md) queries.

{% image
   source="https://docs.dd-static.net/images/bits_ai/business_intelligence.60af363ab2e745f53602c2f96d5df998.png?auto=format&fit=max&w=850 1x, https://docs.dd-static.net/images/bits_ai/business_intelligence.60af363ab2e745f53602c2f96d5df998.png?auto=format&fit=max&w=850&dpr=2 2x"
   alt="Bits Data Analysis chat interface answering a question about top customers driving Electronics growth, alongside an admin overview dashboard showing agent usage metrics and active users" /%}

## Use cases{% #use-cases %}

- Ask ad-hoc business questions without writing SQL.
- Combine business and observability data in a single analysis.
- Save reusable skills for repeated workflows.
- Pull results directly into a [Datadog Notebook](https://docs.datadoghq.com/notebooks.md) for further analysis.
- Visualize results and save them to a [Datadog Dashboard](https://docs.datadoghq.com/dashboards.md) for ongoing monitoring.

## Sample questions{% #sample-questions %}

Ask Bits Data Analysis questions like:

- `What is the ARR for top customers this quarter?`
- `Which products have the most active customers?`
- `What is the monthly support ticket volume trend?`
- `Which products have the highest open ticket count?`
- `Which deals have been in negotiation longer than 30 days?`
- `What is the company headcount by office location?`

## How it works{% #how-it-works %}

1. Type a question in natural language.
1. Bits Data Analysis identifies the relevant tables, writes the SQL, and runs it against your connected data warehouse.
1. Results are returned as a table you can explore.
1. Save the results to a notebook, refine the query, or turn it into a skill.

## Query contexts{% #query-contexts %}

Query contexts are curated instructions per data domain, such as Product or Sales. Each context tells Bits Data Analysis which tables to use, which filters to apply, and which edge cases to avoid. Query contexts are what make Bits Data Analysis accurate and consistent.

Bits Data Analysis can auto-generate query contexts from usage and metadata in your connected BI tools and data warehouses through [Data Observability](https://docs.datadoghq.com/data_observability.md). Data teams then review and refine the generated instructions so anyone on the team can get reliable answers, without analyst expertise or analyst involvement.

### Example: revenue query context{% #example-revenue-query-context %}

A Revenue query context might include instructions such as:

- Use the `fct_subscription_revenue` table as the primary source for ARR and MRR.
- Filter out internal test accounts (`account_type = 'internal'`) by default.
- Treat revenue as recognized in the currency of the parent account, converted to USD using the `fx_rate_daily` table.
- Exclude churned customers from "active customer" counts unless the question explicitly asks about churn.

### Evaluations{% #evaluations %}

Data teams can define evaluations against a query context to verify that Bits Data Analysis answers representative business questions correctly. An eval runs a set of expected question-and-answer pairs against the context. When an answer drifts from the expected result, the eval flags a regression so you can refine instructions before users see incorrect results.

## Skills{% #skills %}

A skill is a saved set of instructions that Bits Data Analysis can execute on demand. Use skills to automate repeated workflows. For example, save a skill that returns weekly support ticket trends by product. You can then run the skill without rephrasing the question each time.

Skills can be kept private, shared with your team, or shared across your company. Sharing lets others run the same workflow without recreating it.

## Permissions{% #permissions %}

Bits Data Analysis only accesses data you have permission to query in your connected data warehouse. To expand your access, use your data warehouse's standard access workflow.

## Accuracy{% #accuracy %}

Bits Data Analysis works well for most questions, but its answers are generated by an AI model. Verify the underlying query and data sources before using results for important decisions, presentations, or external reporting.

## Further reading{% #further-reading %}

- [Build analyses in Datadog Notebooks](https://docs.datadoghq.com/notebooks.md)
- [Datadog Dashboards](https://docs.datadoghq.com/dashboards.md)
- [Query data with DDSQL](https://docs.datadoghq.com/ddsql_editor.md)