> 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/ai-and-automations/agents.md).
# Agents
AI agents that clean your demand history, forecast, find similar products, plan supply and research competitor prices.
Agents are AI processes that work on your data in the background: they clean your demand history, compute the forecast, find similar products and plan your supply. One agent, the Pricing research agent, runs on demand to study competitor prices.
In the sidebar, go to **AI & automations** > **Agents**. Each agent is a card with its description, how often it runs and its next run.
{% hint style="info" %}
On [Flowlity Lite](/flowlity-lite.md), AI agents are a paid add-on and are switched off by default. Ask your Customer Success Manager to activate them.
{% endhint %}
## The agents
* **Demand anomaly agent** (weekly): finds anomalies in your demand history and replaces them with values that reflect normal demand, so the forecast learns from clean data. It handles four types:
* **Outliers**: unusual demand peaks that do not reflect typical customer demand.
* **Shortages**: periods when demand was low because the product was not available. Each site uses one method: **0 stock and 0 demand** (stockout periods with zero demand) or **stock < lot size** (periods when stock below the lot size held sales back).
* **Smooth past promotions**: replaces promotional spikes with the demand expected under normal conditions.
* **Launch detection**: detects when a product moves from its launch phase to a steady demand pattern.
* **Forecast agent** (weekly): computes Flowlity's forecast and keeps learning from your data.
* **Product similarity agent** (weekly): finds the 5 most similar products of each item, from its text, its numbers and its categories. It helps new products with little history get a reliable forecast.
* **Planning agent** (usually daily): recommends order quantities and dates from your forecast and buffers, while respecting your constraints.
* **Pricing research agent** (on demand): compares your price with competitor prices and recommends a strategy.
## How to check what an agent did
{% stepper %}
{% step %}
### Open the agent
In **Agents**, click the agent card.
{% endstep %}
{% step %}
### Read the summary
The **How it works** tab explains what the agent does and sums up its latest results. For the Demand anomaly agent, you see the total number of adjustments and a breakdown by type, with how many were **Accepted**, **Rejected**, **Modified** or **Not treated**. For the Planning agent, you see the order lines it created and their total value.
{% endstep %}
{% step %}
### Act on the corrections
You review each demand correction in the Demand module, on the product itself. See [Investigate and fix demand anomalies](/demand-management/demand/investigate-demand-anomalies.md).
{% endstep %}
{% endstepper %}
## How to check how the Demand anomaly agent is set up
{% stepper %}
{% step %}
### Open the Configuration tab
Click the **Demand anomaly agent** card, then **Configuration**.
{% endstep %}
{% step %}
### Read each anomaly type
Each type shows **Active** or **Inactive**. For **Shortages**, you see which method is used. For each active type, you see whether corrections are accepted automatically or manually: manual corrections wait for your review in the Demand module.
{% endstep %}
{% step %}
### Ask for a change
The configuration is shown for information. Your Customer Success Manager can change it for you.
{% endstep %}
{% endstepper %}
## How to see the constraints the Planning agent respects
{% stepper %}
{% step %}
### Open the Configuration tab
Click the **Planning agent** card, then **Configuration**.
{% endstep %}
{% step %}
### Read the constraints
The list shows each constraint and whether it is **Used** or **Not used** for your site: product constraints (lead time, minimum order quantity, lot size, order frequency), supplier constraints (full truck, supplier MOQ, supplier order frequency), the **Tag family MOQ**, and capacity. The tab also shows the planning horizon.
{% endstep %}
{% endstepper %}
## How to run a competitive price analysis
{% stepper %}
{% step %}
### Open the Pricing research agent
Click the **Pricing research agent** card, or the **Pricing research agent** button in the [Pricing](/revenue-optimization/pricing.md) list.
{% endstep %}
{% step %}
### Choose the agent mode
Under **Agent mode**, pick **Google Shopping** for a quick search, or **Deep stores research** for an in-depth analysis across several stores and sources.
{% endstep %}
{% step %}
### Describe the product
Select a product, then check its **Product name**, **Product description** and **Price**. Under **Area**, choose the **Location** (**France** or **United States**) and the **Search language** (**English** or **French**). With **Google Shopping**, you can also name one **Competitor** to focus on.
{% endstep %}
{% step %}
### Run the agent and download the report
Click **Run agent**. The report gives an executive summary (your price against the market median, the price range, the number of competitors found) and pricing recommendations. Download it with **Download in PDF** or **Download in JPEG**.
{% endstep %}
{% endstepper %}
## Good to know
* **Credits**: each Pricing research run uses AI credits. The remaining balance shows in the agent. Click **Get more credits** to top it up.
* **Similar products**: the Product similarity agent's suggestions show under **Most similar products** when you pick a similar product in the **Product strategy** panel. Its configuration tells you whether products that already have a manual similar link are left out (**Filter products with existing similar link**).
## Related
* [Investigate and fix demand anomalies](/demand-management/demand/investigate-demand-anomalies.md): review the Demand anomaly agent's corrections.
* [Forecast events and demand adjustments](/concepts/concepts/forecast-events.md): how corrections change the demand history.
* [Demand forecasting explained](/concepts/concepts/demand-forecasting.md): how the Forecast and Product similarity agents feed the forecast.
* [Product similarities](/concepts/concepts/product-similarities.md): how similar products are used.
