> 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/consume-with-ai/cookbook/find-a-trustworthy-data-product.md).
# Find a Data Product
**Goal:** You have a business need ("report on quarterly sales revenue") but you do not yet know which Data Product to trust. This recipe takes you from the question to a vetted, healthy product, using only natural language.
**Journey:** Discover → Trust (quality + run history) → confirm ownership.
**Tools used:** `search`, `vulcan_about`, `vulcan_quality`, `vulcan_runs`.
**Example product:** Orders360 (`orders360`).
{% hint style="info" %}
Each step below shows an example prompt as a toggle - phrase it however you like; the assistant works from your intent, not exact wording. Expand a toggle to see the full response. The prompts and responses shown are illustrative examples from a sample catalog; the data products, metrics, and results in your environment will differ.
{% endhint %}
{% stepper %}
{% step %}
### Connect Claude Desktop
Connect Claude Desktop to Data Product MCP so Claude can call DataOS tools on your behalf. Follow [Claude → Claude Desktop](/consume/consume-with-ai/connect-clients/claude.md), then confirm **DataProduct-MCP** appears under **Settings → Developer**.
Confirm you can actually reach products before you start vetting one. Ask what is available — Claude returns Data Products from your DataOS instance, scoped to your token permissions, so you know the connection works and what you are allowed to see.
*Claude Desktop connected to Data Product MCP*
What data products are available to consume?
Here's a summary of the data products available in the catalog, grouped by domain:
***
#### customer-360 (1)
| # | Name | Description | Business Metrics |
| - | ----------------------- | ------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- |
| 1 | Customer Account Health | Customer relationship state, account tier, status, activity, and identity coverage for Helix Industrial | account\_tier\_mix\_pct, active\_customer\_count, customer\_identity\_coverage\_pct (+4 more) |
#### finance (1)
| # | Name | Description | Business Metrics |
| - | ------------------ | -------------------------------------------------------------------------------------------- | ------------------------------------------------------------------ |
| 2 | Receivables Health | Open AR, aging, disputes, promises to pay, and financial distress facts for Helix Industrial | ar\_0\_30\_amount, ar\_31\_60\_amount, ar\_90\_plus\_pct (+7 more) |
#### procurement (1)
| # | Name | Description | Business Metrics |
| - | -------------------- | -------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------- |
| 3 | Supplier Performance | Supplier delivery, fill-rate, rejection, lead-time, and material traceability for Helix Industrial | active\_supplier\_count, incoming\_rejection\_rate, po\_count (+4 more) |
#### sales (1)
| # | Name | Description | Business Metrics |
| - | --------- | ----------------------------------------------------------------- | ----------------------------------------------- |
| 4 | Orders360 | Sales analytics over orders, customers, and products on Snowflake | revenue\_trend, order\_volume, customer\_growth |
*…and more across the `sales_operations`, `logistics`, `quality`, and `manufacturing` domains, scoped to what your token is authorized to see.*
{% hint style="info" %}
For a full example of catalog discovery prompts and responses, see [Discover in natural language](/consume/discover/discover-with-ai.md).
{% endhint %}
{% endstep %}
{% step %}
### Discover candidates for the business question
Start from the need, not the product name. Ask Claude to find Data Products related to your question, then narrow to the ones that expose the exact metric you need.
I need to report on quarterly sales revenue. What Data Products exist for orders and revenue?
Here are the data products most relevant for quarterly sales revenue reporting, grouped by domain:
***
#### sales (2)
| # | Name | Description | Business Metrics |
| - | ----------------------------------------------------- | ---------------------------------------------------------------------------------- | ------------------------------------------------------- |
| 1 | Orders360 (`orders360`) | Sales analytics over orders, customers, and products on Snowflake | **order\_volume**, **revenue\_trend**, customer\_growth |
| 2 | Pricing Performance Analytics (`pricing-performance`) | Pricing effectiveness, contract performance, and sales revenue trends on Snowflake | - |
#### sales\_operations (3)
| # | Name | Description | Business Metrics |
| - | ----------------------------------------------------------- | -------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- |
| 3 | Orders Analytics (`orders-analytics`) | Centralized order lifecycle insights, creation to fulfillment | **daily\_order\_volume**, **daily\_revenue**, daily\_avg\_order\_value |
| 4 | Orders Analytics (Simple) (`orders-analytics-simple`) | Streamlined version on Snowflake for e-commerce/retail | **daily\_order\_volume**, **daily\_revenue**, daily\_avg\_order\_value |
| 5 | **Orders Analytics Platform** (`orders-analytics-platform`) | Bronze → Silver → Gold medallion product with the richest metric set | customer\_lifetime\_value, **daily\_order\_volume**, **daily\_revenue**, revenue\_trend, customer\_acquisition (+10 more) |
***
**Recommendation for quarterly revenue reporting**
* **Orders360** (`orders360`) - clean Snowflake product with `revenue_trend` and `order_volume` directly queryable.
* **Orders Analytics Platform** (`orders-analytics-platform`) - richest metric set including `daily_revenue`, `revenue_trend`, and customer analytics; built on a full medallion architecture.
Would you like me to inspect the semantic schema of any of these to confirm which measures and time dimensions support quarterly slicing?
Which of those has a monthly or quarterly revenue trend metric?
Two distinct data products have a revenue trend metric:
***
#### sales
| # | Name | FQN | Relevant Metrics |
| - | --------- | ----------- | --------------------------------------------------- |
| 1 | Orders360 | `orders360` | **revenue\_trend**, order\_volume, customer\_growth |
#### sales\_operations
| # | Name | FQN | Relevant Metrics |
| - | ------------------------- | --------------------------- | ------------------------------------------------------------------------------------------------------------------- |
| 2 | Orders Analytics Platform | `orders-analytics-platform` | **revenue\_trend** + 14 more metrics including daily\_order\_volume, customer\_lifetime\_value, customer\_retention |
***
**Which is best for quarterly reporting?**
* **Orders360** (`orders360`) is a clean, BI-oriented Snowflake product with `revenue_trend` directly queryable - good for a focused quarterly revenue report.
* **Orders Analytics Platform** (`orders-analytics-platform`) has the broadest metric set and a full medallion architecture, so it's the most production-ready option if you need revenue trend alongside order volume and customer analytics.
Would you like me to inspect the semantic schema of **Orders360** or **Orders Analytics Platform** to confirm the exact time dimensions available for quarterly slicing?
This recipe follows the **Orders360** data product (`orders360`), which exposes the `revenue_trend` metric. The remaining steps vet it before relying on it.
{% endstep %}
{% step %}
### Read the product's identity card
Once orders360 looks like the right candidate, confirm what it is and what it is for. This calls `vulcan_about` and returns the identity card: description, domain, tags, owner, business metrics, use cases, and limitations.
Tell me about the orders360 Data Product - description, domain, tags, and use cases.
Here's the full metadata for the **Orders360** data product:
***
**Orders360** - the Snowflake-backed sales analytics data product
| Field | Value |
| ---------------- | ------------------------------------------------ |
| Description | All information about the orders360 data product |
| Domain | sales |
| Tags | snowflake, orders360 |
| Glossary Terms | glossary.data\_product |
| Owner | dana.lee |
| Business Metrics | revenue\_trend, order\_volume, customer\_growth |
| Use Cases | None configured |
| Limitations | None configured |
| README | Not configured |
***
The product is built on Snowflake and exposes the orders360 lineage and metrics. Use cases and limitations haven't been documented yet. Would you like me to check the semantic schema to see what time dimensions are available for quarterly slicing of `revenue_trend`?
This recipe uses the **Orders360** data product (`orders360`), owned by **dana.lee**.
{% endstep %}
{% step %}
### Check quality
Confirm the data is healthy before you trust a number from it. This calls `vulcan_quality` and returns a dimension-by-dimension summary plus any failing checks.
Are the quality checks on orders360 passing right now? Give me the completeness, validity, uniqueness, and timeliness scores.
Here's the current quality status for the Orders360 data product (as of June 2, 2026 at 00:02 UTC, env: prod):
**Overall status: 🔴 Unhealthy**
| Dimension | Pass | Fail | Warn | Error | Total | Pass Rate |
| ------------ | ---- | ---- | ---- | ----- | ----- | -------------------- |
| Validity | 3 | 1 | 0 | 0 | 4 | 75.0% |
| Completeness | - | - | - | - | - | No checks configured |
| Uniqueness | - | - | - | - | - | No checks configured |
| Timeliness | - | - | - | - | - | No checks configured |
**Failing check**
| Check | Model | Dimension | Outcome |
| ---------------------- | --------------- | --------- | ------- |
| invalid\_email\_format | sales.customers | Validity | ❌ Fail |
**Passing checks**
| Check | Model | Dimension | Outcome |
| ------------------------ | -------------- | --------- | ------- |
| negative\_order\_amounts | sales.orders | Validity | ✅ Pass |
| invalid\_price\_range | sales.products | Validity | ✅ Pass |
| invalid\_rating\_range | sales.products | Validity | ✅ Pass |
**Summary:** Only validity checks are configured (4 checks, 75% passing). Completeness, uniqueness, and timeliness have no checks set up yet. The one failure is `invalid_email_format` on the `customers` model.
Would you like me to drill into why the `invalid_email_format` check is failing, or explore adding checks for the missing dimensions?
{% endstep %}
{% step %}
### Check freshness and run history
Make sure the product refreshed recently and the last run succeeded. This calls `vulcan_runs` and returns the latest run status, timestamp, and rows processed.
When was orders360 last refreshed, and did the last run succeed?
Here's the full run picture for Orders360:
**Last refresh**
The last run succeeded ✅ on June 2, 2026 at 00:00 UTC, completing in 1 min 53 sec across 4 models in prod (`customers`, `orders`, `customer_orders_enriched`, `products`).
**Recent run history (5 runs - 100% success rate)**
| Date | Status | Duration |
| ---------------------- | --------- | -------- |
| Jun 2, 2026 00:00 UTC | ✅ Success | 113.3s |
| Jun 1, 2026 00:00 UTC | ✅ Success | 26.9s |
| May 31, 2026 00:00 UTC | ✅ Success | 17.5s |
| May 30, 2026 00:00 UTC | ✅ Success | 17.4s |
| May 29, 2026 00:01 UTC | ✅ Success | 88.3s |
**Schedule & pending status**
All models run `@daily`. However, 4 models are currently flagged as overdue (next run was expected May 30):
| Model | Status |
| -------------------------- | ---------- |
| customer | ⚠️ Overdue |
| customer\_orders\_enriched | ⚠️ Overdue |
| orders | ⚠️ Overdue |
| products | ⚠️ Overdue |
This may be worth flagging to the owner **dana.lee** for investigation.
{% endstep %}
{% step %}
### Flag the issues to the owner
The trust checks surfaced two things to act on: a failing `invalid_email_format` check and four overdue models. Ask Claude who owns the product so you can follow up. This calls `vulcan_about` and returns the owner contact.
orders360 has a failing email-format check and four overdue models. Who owns it, and can you draft a short note flagging these so I know who to follow up with?
The owner is **dana.lee** (from the metadata we already retrieved). Here's a draft note:
> Hi Dana,
>
> I'm reaching out regarding two open issues on the **Orders360** data product (`orders360`) that require your attention:
>
> **Issue 1 - Quality check failure**
>
> * Check: `invalid_email_format`
> * Model: `sales.customers`
> * Dimension: Validity
> * Status: Failing (overall validity pass rate: 75%)
>
> **Issue 2 - Overdue model schedules**
>
> * Affected models: `customer`, `customer_orders_enriched`, `orders`, `products`
> * Expected cadence: `@daily`
> * Last expected run: 2026-05-30
>
> Please could you investigate and confirm resolution? Let me know if you need any additional context.
The note covers both the email-format failure and the overdue schedule issue with enough detail for the owner to act on it.
{% endstep %}
{% step %}
### Make the trust decision
Ask Claude to summarize the signals into a go / no-go.
Based on quality and freshness, is orders360 safe to use for a quarterly revenue report today?
Based on everything we've already pulled, here's the assessment:
**Orders360 - readiness for quarterly revenue reporting**
| Factor | Status | Detail |
| ------------------- | ---------- | ---------------------------------------------------------------------------- |
| Last refresh | ✅ Good | Ran successfully today, Jun 2 at 00:00 UTC |
| Run reliability | ✅ Good | 100% success rate across last 5 runs |
| Validity checks | ⚠️ Concern | 75% pass rate - `invalid_email_format` failing on `sales.customers` |
| Completeness checks | ❓ Unknown | No checks configured |
| Uniqueness checks | ❓ Unknown | No checks configured |
| Timeliness checks | ❓ Unknown | No checks configured |
| Overdue models | ⚠️ Concern | `customer`, `orders`, `customer_orders_enriched`, `products` flagged overdue |
**Verdict: Proceed with caution ⚠️**
The good news: the data is fresh - it ran successfully this morning and has done so every day this week without failure.
The caveats:
* **Email validity failure** - the `invalid_email_format` check failing on `customers` is unlikely to affect revenue figures directly, but it signals upstream data quality issues that haven't been fully audited. If your revenue report joins on customer identity, there's a risk of mismatched or dropped records.
* **Overdue model schedules** - the four models are flagged overdue (expected run May 30) even though the latest run reported success today. Confirm the schedule status before assuming the data is current.
* **No completeness, uniqueness, or timeliness checks** - there's no automated guarantee that all orders are present, free of duplicates, or within expected time bounds. For a revenue report, these are significant blind spots.
**Recommendation:** Safe to use for a one-off quarterly revenue figure given today's successful refresh, as long as you're not relying on customer email as a join key. Confirm the overdue-schedule flag and flag the gaps to **dana.lee** before treating this as a production-grade source for recurring reporting.
{% endstep %}
{% endstepper %}
## Outcome
You went from a business question to a vetted, named Data Product (orders360, owned by `dana.lee`) without writing any SQL or opening the catalog. More importantly, you know exactly *how far* to trust it: the data is fresh and runs reliably, but a failing validity check, overdue model schedules, and missing completeness checks mean it is safe for a one-off revenue figure - not yet a production-grade source for recurring reporting. You also have a drafted note to flag those gaps to the owner.
This is the difference between an AI that fabricates a confident number and one that hands you a governed answer with its caveats attached.
## Next steps
* [Answer a KPI question](/consume/consume-with-ai/cookbook/answer-a-quarterly-kpi-question.md) - turn this vetted product into a governed, cited answer.
---
# 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/consume-with-ai/cookbook/find-a-trustworthy-data-product.md?ask=&goal=
```
`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.