[Feature] - Collect and categorize pre-defined prompts

[manuw99]

• [x] Description created
• [x] Reviewed by peer:
• [x] Prioritised

Description

We're collecting ideas for pre-defined prompts and then attempting to categorize them in order to better fit the planned drop-down menu style.

Predefined queries could be:

• Generate a process documentation
• What does this process do
• Generate forms for user tasks in the process

Sub-Tasks

```[tasklist] ### Tasks - [x] Create a JSON file with examples of pre-configured prompts - [x] Test and further improve prompts in ChatGPT conversation ```

[manuw99]

Collection of prompts (put into fitting categories):

General Questions:

• What is business process modeling, and why is it important for organizations?
• What does this process do?
• What are common mistakes to avoid when creating BPMN diagrams?
• Can you create a basic bpmn diagram for the following process:

BPMN Help:

• How can I represent decision points in a BPMN Diagram?
• What's the best way to depict parallel activities in BPMN?
• Explain the different types of gateways in BPMN and when to use each.
• Can you provide examples of inclusive and exclusive gateways?
• Can you help me understand when to use manual tasks and automated tasks?
• What are some best practices for naming tasks and activities in BPMN?

Specific tasks:

• Give me ideas on how to represent user tasks in BPMN.
• How do I model a service task that interacts with an external system?
• How can I show that an error occurred during a process step?
• Are there ways to improve this model?

Forms:

• What benefits does the creation and usage of forms have?
• Generate a form for user tasks in the process.
• Include a brief description for each form field in the form.
• How can I optimize this form?
• Design a visually appealing form layout for a process.

Process Documentation:

• Generate a process documentation as text
• Generate a process documentation as a .json file
• Generate a process documentation as a .md
• Generate a process documentation as a .json file using the file process.json as a template
• Generate a process documentation as a .json file using the file process.schema.json as a template
• Generate a process documentation as a .md using the file process.md as a template

[manuw99]

Added JSON file containing the prompts to resources folder 282ead8ddfe722226ec1af15623cda926f7e115f

[manuw99]

I've researched guidelines for creating well written prompts (keyword prompt engineering) and tried to set up a framework for prompts in the BPM domain.

Helpful sources:

• Prompt Engineering Tutorial – Master ChatGPT and LLM Responses https://www.youtube.com/watch?v=_ZvnD73m40o&pp=ygUacHJvbXB0IGVuZ2luZWVyaW5nIGNoYXRncHQ%3D

• Master the Perfect ChatGPT Prompt Formula (in just 8 minutes)! https://www.youtube.com/watch?v=jC4v5AS4RIM

• Master Prompt Engineering: Prompt Recipes - A Framework for Reusable AI Prompts https://www.promptengineering.org/master-prompt-engineering-prompt-recipes/

• Elements of a Prompt https://www.promptingguide.ai/introduction/elements

• What Are Large Language Model (LLM) Agents and Autonomous Agents https://www.promptengineering.org/what-are-large-language-model-llm-agents/

• Asking ChatGPT about well written prompts directly

For our background the following criteria are most relevant:

1. Task: The task that the AI is asked to perform. 2. Instructions: Further instructions if needed (Output, structure, Do's and Don'ts, Examples) 3. Input: (Data) Input or the question. 4. Context: Perspective, goal, target audience 6. (Role/persona): Should the AI take on a role or persona? E.g. "In the role of a BPM-Specialist..." 7. (Tone): As ChatGPT itself said: "Be Polite and Respectful: Maintain a polite and respectful tone throughout your conversation. Remember that the chatbot is a digital entity, but it's always good practice to be courteous." The prompt can be written in a certain tone which will in hindsight influence the tone of the answer. In our case we should stick to a neutral, professional tone. 9. (Type or format of output): The output can be received in a specific format (e.g. JSON, .md) or type.