# Handling High-Risk Users

therappai includes a built-in safety layer that ensures all AI therapy responses remain supportive, calm, and appropriate — even when users express distress. However, **your application** is responsible for how high-risk situations are handled in the user interface.

This guide explains how to responsibly support users who may be experiencing severe stress, emotional overwhelm, or crisis signals.

***

## **What therappai Detects**

Each message sent to therappai is automatically scanned for signs of:

* emotional distress
* self-harm wording
* crisis language
* panic, overwhelming stress, or despair
* harmful or abusive patterns

Even if these signals appear, therappai will still respond with **safe, non-clinical, empathetic language** that helps ground the user.

therappai does **not** diagnose, escalate, or contact emergency services.

***

## **Your Role as the Integrator**

Your application is responsible for:

* how crisis buddy contacts are displayed
* whether additional safety options appear
* which local/national resources you list
* when to encourage the user to reach out for human support
* ensuring the interface does not create a false sense of professional intervention

therappai provides **supportive AI conversations** — not clinical treatment, emergency services, or human supervision.

***

## **Recommended High-Risk UX Patterns**

Below are responsible, widely-accepted ways wellbeing apps handle moments of crisis or high distress.

***

### **1. Show crisis support resources in the UI**

When distress signals appear (detected by your own heuristics, low mood entries, or user-triggered help buttons), show options such as:

* “Talk to someone you trust”
* “Contact your Crisis Buddy”
* “Call your local helpline”
* “Open grounding exercise”
* “Go to safety resources”

This keeps the user empowered and in control.

***

### **2. Present Crisis Buddy options calmly**

Crisis Buddy contacts should appear without judgement or alarmism.

Example:

> **Need support?**\
> You’ve added a trusted contact who might be able to help.

Provide buttons such as:

* “Call now”
* “Message contact”
* “View contact details”

Avoid automated messages or forced actions.

***

### **3. Offer grounding or calming content**

When a user appears overwhelmed, suggest content such as:

* breathing exercises
* grounding tools
* mindfulness practices
* short CBT/DBT skills

This gives users immediate, low-effort support.

***

### **4. Encourage offline help when needed**

Use compassionate messaging such as:

> “I’m really sorry you’re going through this. It might help to talk to someone you trust, or reach out to a local support service if you feel unsafe.”

Keep the tone supportive and non-clinical.

***

### **5. Do not promise professional intervention**

Avoid language that implies:

* clinical assessment
* therapist monitoring
* emergency service contact
* real-time supervision

Clearly state that:

* therappai does not replace professional care
* no emergency intervention is triggered automatically

This protects user trust and sets appropriate expectations.

***

## **What NOT to Do**

To prevent harm or misunderstandings:

#### **Do NOT:**

* call emergency services automatically
* contact Crisis Buddy contacts automatically
* claim clinical oversight or diagnosis
* allow coaches or employers to access private messages
* minimise or dismiss reported distress
* rely solely on AI for crisis intervention

#### **Do NOT show messaging like:**

* “therappai is monitoring your risk level”
* “A therapist has been notified”
* “Emergency services will be contacted”

therappai is **supportive**, not supervisory.

***

## **How High-Risk Support Interacts With Other Features**

High-risk moments can be gently paired with:

#### **Mood Tracking**

If the user logs a very low mood, show:

* grounding tools
* crisis buddy
* support resources

#### **Content Library**

Recommend:

* grounding exercises
* distress tolerance tools
* calming practices

#### **AI Therapy Sessions**

The AI will automatically use a supportive, stabilising tone.

Your UI should reinforce this with safe options.

***

## **Example High-Risk UI Flow**

Below is a common, responsible approach:

1. User expresses distress
2. AI replies with supportive grounding
3. Your app displays a small panel:
   * “Talk to someone you trust”
   * “View your Crisis Buddy”
   * “Try a calming exercise”
4. User chooses what they feel comfortable with
5. App returns to regular features once the user is ready

This keeps control in the user’s hands.

***

## **Summary**

To handle high-risk situations responsibly:

* Use therappai’s built-in supportive responses
* Provide clear, gentle crisis-support options
* Empower users with Crisis Buddy, not automate actions
* Offer grounding content and mindfulness tools
* Never imply clinical monitoring or emergency intervention
* Protect user privacy at all times

therappai gives users a safe, supportive space — your integration determines how that support is surfaced during difficult moments.