> For the complete documentation index, see [llms.txt](https://docs.flowlity.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.flowlity.com/how-tos/investigate-demand-anomalies.md).

# Investigate and fix demand anomalies

This guide walks you through identifying demand anomalies, reviewing the corrections proposed by the Demand Anomaly Agent, and deciding whether to accept or reject them.

{% stepper %}
{% step %}
**Spot anomalies via the Outliers view**

Open the [Demand](/modules/demand.md) module and switch to the **Outliers** view tab. This built-in view surfaces products where the Demand Anomaly Agent has detected unusual demand patterns.

Browse the list and click on a product to open its detail view.
{% endstep %}

{% step %}
**Compare cleaned vs raw demand**

In the product detail chart, toggle on both:

* **Cleaned past demand** — The demand after Flowlity has removed anomalies.
* **Raw past demand** — The original, unmodified demand data.

Periods where the two lines diverge are where the agent proposed corrections. The **Anomaly adjustments impact** row in the data table shows the exact difference per period.
{% endstep %}

{% step %}
**Review agent suggestions**

Go to the [Agents](/modules/agents.md) module and open the **Demand Anomaly Agent** card. The "How it works" tab shows a summary of all recent adjustments, broken down by type:

* **Outliers** — Unusual peaks.
* **Shortages** — Zero-demand periods caused by stockouts.
* **Smooth past promotions** — Promotional spikes smoothed out.

Each adjustment has a status: **Accepted**, **Rejected**, **Not treated**, or **Modified**.
{% endstep %}

{% step %}
**Accept or reject corrections**

For each anomaly, decide whether the agent's correction is appropriate:

* **Accept** — The correction replaces the raw demand in cleaned past demand, improving future forecast accuracy.
* **Reject** — The original raw value is kept. Use this when the anomaly reflects genuine demand (e.g. a real one-time order, not noise).
* **Modify** — Adjust the agent's proposed value to a number you consider more accurate.
  {% endstep %}

{% step %}
**Configure automatic vs manual handling**

In the Agents module, open the Demand Anomaly Agent's **Configuration** tab. You can:

* **Activate or deactivate anomaly types** — Turn off detection for specific types if they aren't relevant.
* **Automatic vs manual** — Set whether corrections are accepted automatically or held for manual review.

{% hint style="info" %}
For products with sparse demand history, the Similar Products agent can help bootstrap a more reliable forecast. Check the product's forecast strategy in [Products settings](/settings/products.md) to see if similar products are configured.
{% endhint %}
{% endstep %}
{% endstepper %}

### Related pages

* [Forecast events and demand adjustments](/concepts/forecast-events.md) — How anomaly corrections and forecast events interact.
* [Demand forecasting explained](/concepts/demand-forecasting.md) — How cleaned demand feeds into the AI model.
* [Demand](/modules/demand.md) — Where anomalies are visible in the product detail view.
* [Agents](/modules/agents.md) — Where to review and configure the Demand Anomaly Agent.


---

# 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:

```
GET https://docs.flowlity.com/how-tos/investigate-demand-anomalies.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.