Skip to main content
All CollectionsGeneral
Chat etiquette: Dos & Don´ts
Chat etiquette: Dos & Don´ts

Set of guidelines when interacting with the chat

Updated over a week ago

Do's for business users utilizing the chat

Provide clear and task-specific prompts

Clarity is key. Ensure that your instructions are specific to the task at hand, which will help the chatbot serve you more efficiently and be succinct.

Use keywords and phrases effectively

Keywords are the compass for our chatbot. They guide it in the right direction to provide you with the best possible answer, so make sure your prompts include relevant terms that are specific to your request.

Opt for structured language

A well-structured prompt in simple language is easily understood by our chatbot, leading to better and quicker responses.

Leverage prompt templates

Make use of predefined prompt templates where applicable to guide the chat’s responses effectively and streamline interactions.

Review chat responses

It is crucial to be aware that our solution is GenAI-based. Therefore, we strongly recommend adopting a "human-in-the-loop" approach by reviewing regularly the chat's responses to ensure they adhere to company guidelines, especially concerning confidential and private data. Look for accuracy, relevance, unwanted biases and any instance where sensitive information might be mishandled.

Report Suspicious Activities

Any unusual activity should be reported immediately. Use the chat responsibly, interactions should always be in line with our company policies and ethical standards.

Be aware of Unique Terms of Use and adhere to your company’s regulations/ policies

When you login for the first time into the chat, you are requested to adhere to Unique´s Terms of Use. These terms apply for all your future interactions with our solution. Ensure that all interactions align with your company policies, data protection guidelines, and ethical standards too in order maintain a secure and trustworthy work environment.

Adhere to Code of Conduct of Azure OpenAI services

Make sure that you adhere to the Code of Coduct when using Microsoft Azure OpenAI services. Please find them here for your reference: https://learn.microsoft.com/en-us/legal/cognitive-services/openai/code-of-conduct?context=%2Fazure%2Fcognitive-services%2Fopenai%2Fcontext%2Fcontext

Don'ts for Business users utilizing the chat

Avoid Ambiguity in Prompts

Ambiguous instructions can lead to errors. Be precise to ensure that the chatbot's data handling is correct and secure.

Don't overcomplicate requests

Complexity can be the enemy of efficiency. Keep your requests to the chat straightforward to avoid confusion.

Avoid sharing sensitive information

Do not input sensitive or confidential information such as client-specific data, personal identifiers, or critical information into the chat. Ensure all interactions comply with our data privacy and security guidelines.

Don’t use jailbreak prompts

Do not use any prompts that may harm the system/service and/or circumvent Large Language Models restrictions. Jailbreak prompts refer to specific set of instructions that attempt to bypass or subvert the intended behavior of the model. These prompts are designed to elicit responses that may violate ethical guidelines, generate harmful content, or exploit the model’s limitations.

Don’t generate and/or execute harmful code

Using the chat to create or execute any code intended to cause harm is prohibited. For example, developing algorithms through the chat with the purpose of initiating fraudulent financial transactions or compromising security systems is not allowed. Finally, make sure to cross-check generated code as it may contain unintended harmful behavior.


Did this answer your question?