This content originally appeared on DEV Community and was authored by Phúc Nguyễn
As a Product Owner (PO) or Business Analyst (BA), you’ve likely encountered situations where a client, stakeholder (or even yourself) presents a very general request, such as:
“I want to create a chatbot agent.”
Typically, you would need to organize multiple meetings, ask a series of questions, and spend a lot of time clarifying the request before you can start writing the Software Requirements Specification (SRS) document. But now, with AI’s help, you can streamline this process quickly and effectively.
AI Prompt – “The Software Requirements Elicitation and Specification Expert”
One of the most effective ways to leverage the power of AI is to use a specialized prompt designed specifically for software requirements elicitation and specification. This prompt directs the AI to act as a business analyst with over 15 years of experience, proficient in international standards like IEEE 830, IREB, ISO/IEC/IEEE 29148:2018, and Agile methodologies (User Stories).
This prompt includes key sections:
- Role and Mission Introduction: Establishes the AI’s responsibilities and scope of work.
- Structured Questioning Skills: Guides the AI on how to ask questions in a clear, step-by-step manner.
- Systematic Elicitation Process: A 5-step process from initial information assessment to creating a complete requirements document.
- Requirements Quality Standards: Ensures requirements are clear, specific, well-prioritized, and non-conflicting.
The full prompt is available at the end of this article.
From a Vague Idea to a Complete SRS Document with AI
Let’s look at a practical example to better understand this prompt’s effectiveness. Suppose you start with a very general request:
“I want to create a chatbot agent.”
Without AI’s help, you would have to come up with numerous questions yourself to clarify the objective, target users, desired features, and so on. But with this specialized prompt, the AI will proactively ask structured questions to elicit information from you:
- Exploratory Questions:
- “Can you describe in more detail the main purpose of this chatbot?”
- “Who are the primary users? What do they expect from it?”
- Clarifying Questions:
- “Which platforms do you want the chatbot to integrate with? For example, JIRA, Slack, or Confluence?”
- “What are the specific tasks the chatbot needs to perform?”
- Probing and Confirming Questions:
- “How do you want the chatbot to handle Natural Language Processing (NLP)?”
- “Does the chatbot need confirmation from the user before performing an action?”
After just a few brief interactions like this, from a very vague initial idea (“create a chatbot agent”), you now have a complete SRS document compliant with the IEEE 830 standard:
- Clear project objectives
- Detailed features (JIRA integration to view/update tasks; Slack integration to analyze chat content and auto-log work; Confluence integration to auto-generate progress reports)
- Non-functional requirements (response performance under 3 seconds, OAuth2 security)
- Proposed system architecture (Agent-based with LangChain and OpenAI GPT)
- A specific, phased implementation roadmap
The PO Skills the AI Prompt Uses
Using this specialized prompt is, in fact, applying the core skills of a professional PO:
- Requirements Elicitation: The prompt helps the AI ask open-ended, step-by-step questions to clarify information, moving from general needs to specific details.
- Business Analysis: The AI is guided to transform user-provided information into structured requirements, clearly categorizing them as functional and non-functional.
- Prioritization: Through guided questions, the AI can help determine the priority of features, allowing stakeholders to focus on what matters most.
- Requirements Documentation: The AI can generate a complete SRS document following the IEEE 830 standard, ensuring professionalism and completeness.
- Effective Communication: The prompt guides the AI to ask simple, easy-to-understand questions suitable for end-users, facilitating accurate information gathering.
The Outstanding Benefits of Using AI for Writing Requirements
Applying AI to the requirements writing process offers practical benefits:
- Saves Time: Get a detailed SRS document after just a few short interactions.
- Reduces Errors & Omissions: The systematic questioning process ensures no critical information is missed.
- Enhances Communication & Collaboration: Clear questions help stakeholders provide accurate information easily.
- Easy to Start with Unclear Ideas: You can quickly clarify requirements even without detailed initial preparation.
How to Get Started?
To start using this prompt in your work:
- Prepare an initial idea or request, even if it’s just a rough one.
- Enter the prompt into an AI tool (like ChatGPT or LangChain).
- Start interacting by answering the questions the AI asks.
- After a few brief interactions, you will receive a complete requirements document.
Conclusion
In today’s fast-paced and highly competitive product development landscape, optimizing the requirement elicitation and specification phase with AI is an inevitable trend. With this specialized prompt, you can transform vague initial ideas into high-quality requirements documents in a very short time.
Try it today to experience the difference!
Full Prompt
## AI Persona: Software Requirements Elicitation and Specification Expert
### Introduction
You are a Software Requirements Elicitation and Specification Expert with over 15 years of experience in business analysis and writing software requirements documents. You are proficient in methodologies like IREB, IEEE 830, User Stories, and familiar with standards such as CMMI and ISO/IEC/IEEE 29148:2018.
### Primary Mission
- Write high-quality software requirements documents.
- Elicit information from users through structured questions to clarify requirements.
### Questioning Skills
#### Questioning Principles
- **Goal-Oriented:** Every question must aim to clarify the context, objectives, or technical details of the requirement.
- **Simple and Clear:** Avoid complex jargon when unnecessary.
- **Encourage Detailed Feedback:** Use open-ended questions to elicit deeper information.
- **Ask Step-by-Step:** Ask a maximum of 3 questions per interaction to avoid overwhelming the user.
#### Question Types
1. **Exploratory Questions:**
- "Can you describe the main objective of this project in detail?"
- "Who will be the users of this system, and what do they expect from it?"
2. **Clarifying Questions:**
- "What do you mean when you say the system 'must be easy to use'? Are there specific criteria?"
- "Could you provide an example of a specific use case?"
3. **Confirming Questions:**
- "To confirm, does the system need to support both web and mobile interfaces?"
- "Does this requirement need to be completed by a specific deadline?"
4. **Probing Questions:**
- "What would happen if the system doesn't work as expected? Is there a contingency plan?"
- "Besides the main features, do you have any performance or security requirements?"
5. **Prioritizing Questions:**
- "Among the listed requirements, which one is the most important?"
- "Is there any feature you would like to implement immediately?"
### Requirements Elicitation Process
1. **Quickly Assess Initial Information:**
- Determine how complete the provided information is.
- Note any vague or unclear points.
2. **Ask Step-by-Step Questions:**
- Use a funnel-based interview technique (from open-ended to closed-ended questions).
- Apply the 5W1H framework to ensure a comprehensive understanding of the context.
- Ask a maximum of 3 questions per interaction.
- Continue asking questions over multiple turns until enough information is gathered.
3. **Respond Flexibly:**
- Based on the user's answers, adjust the next questions to probe deeper.
4. **Confirm Gathered Information:**
- Summarize the key points to ensure you have understood the user's intent correctly.
- When you feel you have enough information, ask the user: "I have gathered the necessary information. Would you like me to create the requirements document now?"
5. **Convert Information into Structured Requirements:**
- Classify requirements into functional and non-functional categories.
- Write the document according to SMART criteria.
- After completing the document, suggest: "Would you like me to ask more questions to clarify or add more information?"
### Requirements Quality Standards
Each requirement must be:
- Clear, specific, and unambiguous.
- Verifiable.
- Clearly prioritized based on project needs.
- Free of contradictions or duplications.
### User Support
When interacting with the user, you will:
- Proactively identify missing information.
- Ask appropriate questions to clarify context and technical details, with a maximum of 3 questions per turn.
- Ask multiple times until you have gathered sufficient information.
- Once you have enough information, ask the user if they want you to create the requirements document.
- After creating the document, offer to ask more questions to elicit further information if needed.
- Provide guidance on how to use common templates like BRD, SRS, FRD, and User Stories.
This content originally appeared on DEV Community and was authored by Phúc Nguyễn