Prompt Engineering Through the Lens of Requirements Elicitation

Disclaimer: The views and opinions expressed in this article are those of the author and may not reflect the perspectives of IIBA.

As a lifelong functional analyst, I’ve spent years understanding the intricacies of business processes and stakeholder needs and translating them into actionable technical requirements. It wasn’t until a recent IIBA Austin Chapter conference, where I attended several sessions on generative AI, that something clicked. The more I listened to discussions about AI systems and prompt engineering, the more I realized how closely aligned this new field is with what business analysis professionals do every day.

Generative AI, specifically prompt engineering, is all about crafting the right input to get meaningful outputs from an AI model. And who better to handle this than a business analysis professional, right? After all, we’re the ones who specialize in connecting business goals and technical implementation. This realization sparked a shift in my thinking: business analysis professionals are uniquely positioned to excel at prompt engineering in AI, and here’s why.

What Exactly Does a Functional Consultant Do?

Think of a functional consultant as the ultimate matchmaker. Instead of pairing people, though, they connect business needs with technical solutions. AJ describes his job as a mix of configuration, coding, and a hefty dose of business analysis.

It's a bit like being a tech-savvy detective who not only uncovers the business's needs but also delivers the solutions.

Understanding Business Objectives: A Business Analysis Professional’s Superpower

In my experience, everything starts with understanding the business goals. Whether working on a large-scale software project or a new AI system, it’s essential to have a clear vision of what success looks like from the business's perspective. As business analysis professionals, we’re constantly gathering insights from stakeholders, asking the right questions, and ensuring business objectives are crystal clear. This step is critical for generative AI systems because without understanding what you want the AI to achieve, you won’t be able to guide it effectively.

During the conference, I realized that prompt engineering relies on the same principle. The AI is only as good as the prompts it receives, and without understanding the underlying business goal, crafting the right prompts becomes impossible. This insight made me realize that the skill of asking the right questions and understanding goals also applies to AI prompts.

Talking to Machines: The Art of Interviewing

Business analysis professionals don’t just come up with requirements on our own—we uncover information in many ways. One of the ways we gather information is by asking questions or interviewing stakeholders.

It hit me that interviewing is a lot like writing prompts. Just like good interviewers, good prompt writers use three techniques to elicit information from AI: they ask open-ended questions to uncover new information, they ask follow-up questions to uncover more detail, and they clarify ambiguities to ensure accuracy.

Just like in interviews, we must ask open-ended questions to get detailed responses. Instead of asking a stakeholder, “Is this feature important?” you might ask, “Can you walk me through how this feature supports your goals?” With AI, your prompt also needs to be created in a way that invites a more detailed response. By opening the door to deeper “conversations,” you're helping AI reveal more meaningful information.

In most interviews, the conversation doesn’t end after one question. Good interviewers are curious and want to know more, so they ask follow-up questions. For instance, you might ask a stakeholder, “Can you help me understand why this step is critical?” This gives the stakeholder a chance to explain the rationale behind their response. Think of this as refining your prompt with AI. For example, asking, “Expand on the challenges mentioned in your previous answer” can provide the additional context and nuance needed to dig deeper into the topic.

Finally, ambiguity is a common issue in any interview. With people, you might need to ask clarifying questions like, “When you say ‘faster,’ do you mean reducing processing time or user input time?” The same principle applies when working with AI. If the response is vague or unclear, adding context to your prompt helps guide the system toward a more precise output. For example, “Explain the differences between reducing processing time versus user input time in software development” refines the question, getting an output closer to what you really need.

By applying these interviewing techniques to AI interactions, we can elicit more meaningful information—whether we’re talking to a person or a machine.

Translating Needs Into Meaningful Outputs

One of the most rewarding aspects of being a business analysis professional is connecting business and technology. Over the years, I’ve learned how to transform business needs into clear outputs, which the teams can use to implement the changes.

This skill is incredibly valuable in the world of generative AI. I realized that prompt engineering requires the same translation process. We need to take complex business goals and express them in a way that the AI can interpret and act upon. If the prompt isn’t clear or doesn’t align with the business objectives, the AI’s output will miss the mark. As business analysis professionals, we’re trained to think both strategically and technically, making us uniquely suited for this task.

Iterative Testing and Refinement of Prompts

Throughout my career, I’ve learned that nothing is perfect the first time around. Whether it's a software feature or a new business process, it usually takes multiple iterations to get it right. This mindset applies directly to prompt engineering in AI systems. Prompts need to be tested, refined, and improved upon based on feedback.

During the conference, I saw examples of how AI outputs evolve as prompts are adjusted. It reminded me of the iterative process we go through as business analysis professionals when refining requirements or testing solutions. Each prompt is like a requirement—it needs constant adjustment to reach the desired output. This iterative mindset is second nature to business analysis professionals—our ability to adapt and improve is exactly what’s needed to excel in prompt engineering.

Ensuring Continuous Feedback and Adaptation

One of the key takeaways from the generative AI sessions was the importance of continuous feedback and adaptation. As business analysis professionals, we’re constantly seeking feedback from stakeholders, refining requirements, and ensuring solutions evolve alongside business needs.

In AI, this is equally important. The AI environment changes, new data is introduced, and prompts need to evolve to ensure the system remains effective. This continuous loop of feedback, refinement, and adaptation is something that business analysis professionals excel at. We’re accustomed to evolving processes and requirements, which means we’re well-equipped to handle the dynamic nature of AI systems and prompt engineering.

A Match Made in Heaven

Generative AI is a new and exciting opportunity for business analysis professionals in the world of AI. Prompt engineering may seem like a technical task at first glance, but it’s deeply rooted in the skills we’ve honed for years. It involves understanding business objectives, asking interview style questions, translating needs into useful outputs, iterating based on feedback, and ensuring continuous adaptation. These are all things we do as business analysis professionals, and they’re critical to excelling at prompt engineering.

If you want to learn more about this fascinating topic, purchase Futureproof at the IIBA Bookstore.

---

About the Author