Just like humans, AskX performs better with well-organized documents.
Semantic structure through headings helps understand the information, and that works for AI too. For optimal results, try using semantic headings (H1/H2/H3/H4).
Asking instead of searching is something we all need to build habit on.
If you type just keywords, AskX will reformulate it as a question, but you may not get the most relevant information. By formulating your query as a question, you help AskX think critically about what it should look for.
AskX relies on Search to give you answers. If you know a specific term must appear in the results, you can boost its importance by enclosing it in double quotes: What is "AskX"
.
By using double quotes, you signal to the system that these words are particularly significant for your query, often leading to more precise and relevant answers.
This is a counter intuitive tip: often, you shouldn't use the keywords most present in your data, like "{your company name}".
The keyword "{your company name}" is everywhere in your data. Since AskX relies on Search for one its steps, using this keyword might add a lot of noise.
You can specify instructions in AskX on how you want the answer to be displayed.
But as a shortcut you can simply write these instructions after the ? mark:
"What is Slite? generate a blog article about it."