====AI Models====

You can use your capture data to train AI predictive models that can estimate your resulting values based on a chosen set of **inputs**. 

For example, you might estimate the minimum and maximum tolerances of a system based the various different types of product you produce, as well as ambient environmental conditions.

You can access this page from [[the_capture_list|the capture list]] by opening **Search** and choosing **AI Model**.

{{aiquery.png}}

===Using===

First, choose the specific AI model you'd like to use.

A single type of Capture may have more than one type of AI model connected to it. Each model is designed to answer a specific type of question that allows you to predict or optimise your process.

When you choose a particular model, you then simply enter the parameters you know about in the fields provided, and press **Query AI**.

//Note: If you leave one or more fields empty, the system will use the **median value** of that field from the sample data.//

===Examples===

Our [[Bakery Demo|https://demo.optrix.com.au/s/default]] contains two different AI models based on the same data.

One gives the //suggested system setup// to achieve good results, while the other takes your current machine settings and suggests what the quality feedback values - in this case, the **rise level** and **brownness** of the cupcakes - would be.