> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cognigy.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Thresholds

> Set the score boundaries that decide if Cognigy NLU recognizes an intent, asks for reconfirmation, or rejects it. Learn how to configure thresholds.

The *thresholds* feature provides configurable score boundaries
that determine how the NLU model interprets and classifies user inputs based on their intent.
This feature enables precise control over intent recognition,
ensuring the NLU can confidently identify intents, prompt for reconfirmation when needed, or reject unclear inputs.

## Key benefits

* **Improved accuracy**. Adjustable thresholds enhance the reliability of intent mapping by setting clear confidence levels.
* **Flexible control**. You can fine-tune the system's sensitivity to match specific use cases or performance needs.

## Prerequisites

Before using thresholds, ensure the following:

* You have an active flow in the project.
* You've defined intents so the NLU can classify user inputs.

## How to use

Explore threshold types, then set them up.

### Explore threshold types

| Type | Description |
| - | - |
| **Confidence Threshold** | Intents with a score higher than this threshold are considered confirmed. |
| **Reconfirmation Threshold** | Intents with a score higher than the reconfirmation threshold but lower than the confidence threshold require reconfirmation. |
| Not found threshold | Intents with a score lower than the reconfirmation threshold aren't considered. |

#### Reconfirmation threshold specifics

The reconfirmation threshold behaves differently depending on whether a [confirmation sentence](https://docs.cognigy.com/ai/platform-features/nlu/intents/overview#apply-additional-settings) exists in your intent:

* With a confirmation sentence, the reconfirmation threshold triggers the confirmation sentence.
* Without a confirmation sentence, the reconfirmation threshold confirms the user input.

### Set up thresholds

You can configure thresholds at either the flow or project level. To set them up:

> [tab] Flow level
>
> 1. In the navigation menu, go to **Design > Flows** and open the flow.
> 2. Go to the **Settings** tab, then select **Configuration**.
> 3. Activate **Thresholds**. As soon as you activate it, this setting overrides the threshold values set up at the project level.
> 4. Adjust the handles of the **Intent Threshold** slider:
>    * **Reconfirmation Threshold** (default: 0.2):
>      * Move the handle left to increase reconfirmation prompts.
>      * Move the handle right to reduce reconfirmation prompts.
>    * **Confidence Threshold** (default: 0.4):
>      * Move the handle left to allow lower-confidence matches.
>      * Move the handle right to require higher confidence for matches.
> 5. Click **Save changes**.

> [tab] Project level
>
> 1. In the navigation menu, go to **Manage > Settings**.
> 2. On the **Settings** page, click **NLU Settings**.
> 3. In the **Thresholds** section, adjust the threshold sliders:
>    * **Reconfirmation Threshold** (default: 0.2):
>      * Move the slider left to increase reconfirmation prompts.
>      * Move the slider right to reduce reconfirmation prompts.
>    * **Confidence Threshold** (default: 0.4):
>      * Move the slider left to allow lower-confidence matches.
>      * Move the slider right to require higher confidence for matches.
>
> Fine-tune the sliders: move them closer for decisive responses or farther apart for cautious reconfirmation.

## Example

Imagine a user asks `Can you book a flight?` in the chat with the agent. The NLU model assigns a score to the `BookFlight` intent:

| Score | Threshold range | Outcome | Agent response |
| - | - | - | - |
| 0.5 | Above confidence threshold (0.4) | The intent is confirmed and the conversation proceeds. | None |
| 0.3 | Between reconfirmation threshold (0.2) and confidence threshold (0.4) | The agent seeks confirmation. | `Did you mean to book a flight?` |
| 0.1 | Below reconfirmation threshold (0.2) | The intent isn't recognized. | `I'm not sure what you mean.` |

## More information

* [Intents](https://docs.cognigy.com/ai/platform-features/nlu/intents/overview)
