> For the complete documentation index, see [llms.txt](https://v2.dataos.info/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://v2.dataos.info/consume/v1/consume-with-ai/understand-in-natural-language.md).

# Understand in natural language

Before you ask a business question, it pays to know what a product can answer, whether the data is fresh and healthy, and where the numbers come from. Your [AI assistant](/consume/v1/consume-with-ai/connect-a-client.md) pulls all of that without opening the catalog, [Studio](/consume/v1/evaluate/query-in-studio.md), or a [SQL client](/consume/v1/activate/database-clients.md). This is the AI mode of [Evaluate](/consume/v1/evaluate/overview.md).

Responses are illustrative and depend on your catalog and permissions. Five things you can check, with the questions that get them:

## What is queryable

The product's queryable surface, before you ask for numbers: measures, dimensions, time dimensions, segments, metrics, and semantic joins. Treat it as the menu, not the order.

<table><thead><tr><th width="219.6915283203125">Ask about</th><th>Example</th></tr></thead><tbody><tr><td>Available metrics</td><td><code>What metrics does orders360 expose?</code></td></tr><tr><td>Dimensions for analysis</td><td><code>What dimensions can I use for revenue analysis?</code></td></tr><tr><td>Fields in a model</td><td><code>What fields does the orders_enriched model have?</code></td></tr><tr><td>Capability check</td><td><code>Can this product answer questions by customer segment?</code></td></tr></tbody></table>

## Lineage

Where the data comes from and what would break if a source changed: upstream sources, downstream dependents, transformation hints, an impact summary, sensitive tables, and owners to notify. It needs an exact product or table name (search first if unsure), and it returns no row values.

| Ask about         | Example                                                 |
| ----------------- | ------------------------------------------------------- |
| Upstream sources  | `Where does the revenue figure in orders360 come from?` |
| Downstream impact | `What would break if I retire this table?`              |
| Sensitive data    | `Does this table's lineage touch any PII columns?`      |

## Table profile

Whether a table is usable: row count, size, last-profiled time, null rates, distinct counts, min, max, and distributions for queried columns. It returns the latest profile (it does not re-run the profiler) and no row contents.

| Ask about          | Example                                        |
| ------------------ | ---------------------------------------------- |
| Freshness          | `How fresh is the orders_enriched table?`      |
| Null columns       | `Which columns in orders_enriched have nulls?` |
| Value distribution | `What is the distribution of order_status?`    |

## Quality

Whether configured checks are passing: a dimension-by-dimension summary (completeness, validity, uniqueness, and any other dimensions the product evaluates), the last evaluation time, failing checks with model and column, and diagnostics for a specific check. It reads the latest state and does not modify or retrigger checks.

<table><thead><tr><th width="177.31414794921875">Ask about</th><th>Example</th></tr></thead><tbody><tr><td>Overall quality</td><td><code>Are the quality checks on orders360 passing?</code></td></tr><tr><td>Failing checks</td><td><code>What is failing on the orders Data Product right now?</code></td></tr><tr><td>Check detail</td><td><code>Why is the invalid_emails check failing?</code></td></tr></tbody></table>

## Run history

Whether the pipeline ran and when data was last refreshed: latest run status, timestamp, rows processed, models affected, duration, errors, and the next scheduled run. It does not retrigger or schedule runs.

<table><thead><tr><th width="210.256591796875">Ask about</th><th>Example</th></tr></thead><tbody><tr><td>Latest run</td><td><code>When was orders360 last refreshed?</code></td></tr><tr><td>Run errors</td><td><code>Why did the last run fail?</code></td></tr><tr><td>Recent history</td><td><code>Show me the last five runs for this product.</code></td></tr></tbody></table>

## A typical trust flow

Check the **schema** to see what is on offer and **quality** to confirm checks pass. Check **runs** to confirm a recent refresh. Check **lineage** when you need to know where a number comes from, and the **table profile** when you want to check nulls or distributions before querying. When the product checks out, [activate](/consume/v1/activate/overview.md) it, or ask a question directly: see the [Answering KPI Questions in Natural Language](/consume/v1/recipes/answer-a-kpi-question.md) recipe.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://v2.dataos.info/consume/v1/consume-with-ai/understand-in-natural-language.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.