# Handling errors

> When talking about chatbots and the mistakes they make, it’s always talked about handling them ‘gracefully’. However, in industrial situations, they must also be handled quickly so users can move on and possibly get another support source. There are often time-sensitive actions within industrial situations that require chatbots and workers to move quickly to find solutions. If a chatbot response is inaccurate or the chatbot cannot give users the solution they need, it can be consequential and even costly.

# 

## Examples 

![Handling errors example apology](https://www.figma.com/design/wEptRgAezDU1z80Cn3eZ0o/iX-Pattern-Illustrations?node-id=3218-4344&t=etx1DcSbA7VDx5xD-4)

Here the chatbot is immediately apologetic and tells the user they cannot complete the task. It’s efficient but lacks empathy and fails to give the user any kind of alternative support. 

![Handling errors example apology and explanation](https://www.figma.com/design/wEptRgAezDU1z80Cn3eZ0o/iX-Pattern-Illustrations?node-id=3218-4376&t=etx1DcSbA7VDx5xD-4)

Here, although it’s clear what the user wants, they can’t give the chatbot the information it needs to fulfill the request. In this situation, the chatbot immediately says they cannot carry out the task, explains why, and then tells the user where to find a solution. We recommend this three-step approach for chatbot responses and for error messages in general.   
## Dos and Don’ts

  
    
      Do assume your chatbot makes mistakes
      Do try to ‘break’ your chatbot in testing
      Do train your chatbot to tell users immediately if they cannot complete a task
    
  
  
    
      Don’t forget to add a disclaimer stating that information could be incorrect