<aside>
On this page:
</aside>
<aside>
🧠 Product Docs
🤖 Prompting Tips
💻 Use Cases
📚 Resources
🍄 About Expanse
</aside>
Prompting is any message you send to an AI. The quality of your prompt determines the quality of your output. AI models have proven very sensitive to small changes in prompts, so it pays to experiment a lot, and a small tweak may greatly improve your results. Therefore, it is really important to develop good prompting styles/habits to get valuable outputs from the AI.
Roles define the AI's expertise for your specific needs. Using Roles gives the AI context to what you are working on, reducing the amount of off-topic ideas it might have. Chatting while using a Role ensures that the AI’s responses remain relevant and specialized to your needs.
There are a number of tasks that you are likely to repeat in your everyday workflows. Saving these prompts to your Prompt Library, will allow you to work faster and save you from typing the same thing over and over again.
Here are some examples of common Prompts:
While the subject/content of these tasks may change, because the underlying purpose of the prompt is the same, it pays to spend time crafting the prompt to produce relevant outputs. A good start would be to use Expanse’s Magic Pen feature which will give you a more structured and detailed prompt to start. But it’s worth noting that to truly create consistently valuable responses you should spend time adding details and refining the prompt as you use it.
Examples of details you might want to add to your prompt:
<aside> đź’ˇ
Here is an example of a Prompt with specified details on the style and output of the response
To get the most relevant responses from an AI, your prompt should be as specific as possible. So when creating a Role or Prompt with the Magic Pen tool, try to add context-specific words.
For example, rather than specifying a “super duper amazing chef” you should add specific details that you want to bring out in the LLM’s response, such as “pastry chef specializing in French patisserie and intricate chocolate sculptures”.
<aside> đź’ˇ
Using Expanse’s Magic Pen Tool is a great way to generate a well-structured Role, remember to include your own Guidelines to make it more personalized to your needs.
Just like human conversations, AI chats work best when kept fresh and focused. Long conversations can cause the AI to forget earlier context and slow down response times. To keep conversations efficient and relevant, start a new Thread every ~20 messages - the more frequent, the better.
Starting new Threads is a great habit, but retaining context can be a challenge. One solution is to ask the AI for a summary at the end of a chat, then refine and reuse it in future conversations. Simply ask the AI to summarize your discussion, copy that summary to a fresh thread, and say "Let's continue from here." See the prompt below.
<aside> đź’ˇ
Save this prompt to use whenever you need to summarize a Thread
Sharing specific parts of files directly related to your current task helps the AI focus on what's important and prevents from processing unnecessary data and therefore reduces the risk of it returning irrelevant or incorrect responses. Smaller files also save tokens and Credits, while reducing the amount of time it takes for the AI to respond.
To get the most effective answers, only upload files that contain essential information/data the AI needs to answer.
Examples of relevant files to upload:
