> For the complete documentation index, see [llms.txt](https://help.jurisphere.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.jurisphere.ai/resources/prompting-guide.md).

# Prompting Guide

## Anatomy of a Prompt

{% stepper %}
{% step %}

### Role-definition

Explain the tool's role for the specific task. For example, "*You are assisting a law firm partner in performing a title deed due diligence*."
{% endstep %}

{% step %}

### Context

Set out the facts surrounding the task that you require the tool to perform. For example, "*The parties have been in dispute for over 20 years, and have now reached a settlement after mediation attempts.*"
{% endstep %}

{% step %}

### Instructions

Direct the exact task that you require the tool to do. For example, "*I need an analysis of the following features of these documents - indemnity, change of control, termination and potential red flags.*"

{% hint style="info" %}
You can expand and explain each part of the task in detail. Remember, <mark style="color:green;">**more specific your instruction, more accurate the response**</mark>!
{% endhint %}
{% endstep %}

{% step %}

### Formatting

Describe the format - **tables, lists, paragraphs, flowcharts**, etc. You can even specify exact column headings, paragraph lengths, <mark style="color:orange;">colours for font</mark>, <mark style="background-color:purple;">shading</mark>, etc.&#x20;

This is also an optimal portion to provide templated language or reference.
{% endstep %}

{% step %}

### Output

In the final part, you can mention what the output will be used for. For example, "The output note of arguments will be tendered in court on behalf of the petitioners."
{% endstep %}
{% endstepper %}

{% hint style="info" %}
While the tool excels at summarization, analysis, and generating insights, it’s important to

* Review outputs carefully for accuracy - particularly for tasks involving voluminous documents.
* Understand that legal nuances may require advanced human thinking and judgment.
  {% endhint %}

## Effective Prompting Techniques

<table><thead><tr><th width="131.069580078125"></th><th valign="top">To Follow</th><th valign="top"> To Avoid</th><th valign="top"> Accessible and Inclusive Considerations</th></tr></thead><tbody><tr><td>Use the document<br>scope wisely</td><td valign="top">Select only the<br><strong>documents you need</strong> for<br>your task. You are smarter than AI in deciding relevance.</td><td valign="top">Don’t expect the tool to<br>access external or online<br>sources, or leave out irrelevant sets of documents.</td><td valign="top">For research,<br>transition to the<br>research tool by<br>selecting relevant text<br>from the responses in<br>Document Review.</td></tr><tr><td>Be clear and<br>specific</td><td valign="top">Use targeted instructions<br>like “<em>Explain the</em><br><em>obligations section of</em><br><em>this agreement from the buyer's perspective</em>.”</td><td valign="top">Avoid overly broad<br>prompts like “<em>Summarize</em><br><em>this agreement</em>.”</td><td valign="top"><strong>Break complex tasks</strong><br><strong>into smaller</strong>,<br>manageable ones for<br>better precision.</td></tr><tr><td>Include context</td><td valign="top">Specify the scope of the<br>task, e.g., “<em>Party ABC has breached the contracted dated 01.02.2000 resulting in severe costs of XYZ. Analyze the</em><br><em>indemnity clause in</em><br><em>relation to the contract</em>.”</td><td valign="top">Don’t assume the tool<br>will intuitively focus on<br>the sections you care<br>about without guidance.</td><td valign="top"><strong>Explaining what task</strong><br><strong>you are doing</strong>, or what<br>the output is being<br>used for can improve<br>the tool's understanding<br>of your needs.</td></tr><tr><td>Guide the format</td><td valign="top">Specify output<br>preferences, e.g.,<br>“<em>Create a table of ....</em>” or<br>“<em>Prepare a</em><br><em>comprehensive note with 6 sentences in each</em><br><em>paragraph ....</em>”</td><td valign="top">Avoid ambiguous<br>instructions like “Make it detailed and<br>professional.”</td><td valign="top">Include desired formats<br>(e.g., <code>paragraphs</code>,<br><kbd>bullet points</kbd>, <strong>text</strong><br><em>formatting</em>, <mark style="color:purple;">colours</mark>) in<br>your prompts to shape<br>the output.</td></tr><tr><td>Rephrase and<br>refine</td><td valign="top">Iterate and clarify when<br>tackling nuanced issues,<br>by <strong>refining your query in</strong><br><strong>the follow-ups</strong> for deeper<br>insights.</td><td valign="top">Don’t assume the AI can<br>provide perfect answers<br>without follow-up<br>prompts.</td><td valign="top">Tools start with a broad<br>and general approach,<br>but are very receptive<br>to instructions.</td></tr></tbody></table>

## Optimal Usage Tips

<table data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><strong>Document Scope</strong></td><td>The tool reviews only the<br>documents you select - choose wisely, and guide using prompts.</td></tr><tr><td><strong>Input Limits</strong></td><td>For tasks involving very<br>lengthy documents, focus<br>on specific sections or use multiple prompts for better results.</td></tr><tr><td><strong>Output Constraints</strong></td><td>If you need to generate lengthy drafts, break it down section-wise into multiple prompts and set specific word limits.</td></tr></tbody></table>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://help.jurisphere.ai/resources/prompting-guide.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
