For over 20 years, I’ve worked as a business analyst, solving problems by uncovering root causes, defining solutions, and structuring clear requirements for change. The more I’ve explored AI and prompt engineering, the more I’ve realised—this skillset translates perfectly.

If you’ve ever felt frustrated by AI’s limitations, it’s probably because you’re not approaching it with an analyst’s mindset. Let’s fix that.

Why Business Analysis and Prompt Engineering Are So Similar

At its core, business analysis is about solving problems strategically. My career has been built on this approach:

1. Uncover the real issue – Not just surface symptoms, but the root cause with it’s issues and risks.
2. Outline requirements for success – Including costs, assumptions, and timelines.
3. Design a solution – One that’s realistic, actionable, and effective.

And across all of this, brainstorm, refine and iterate until you’ve hit the mark.

The exact same principles apply to AI prompt engineering.

When people get poor AI responses on large language models like Chata GPT, it’s usually because they’ve skipped these steps. They haven’t defined the problem, constraints, or expected structure, so the AI is left guessing. And just like a business analyst doesn’t define solutions alone, great prompt engineering isn’t a one-shot process.

Brainstorming with AI is like brainstorming with a project team.

In a real-world project, we don’t walk into a meeting, throw out a question, and instantly get a perfect answer. We iterate. We refine. We challenge assumptions. Working with AI is the same. Your first prompt is rarely the final draft—it’s a conversation, not a one-and-done request.

Let’s break it down with a real example.

How an Analyst Would Approach AI Prompting

Let’s say you need AI to generate a business proposal for a new AI-powered customer service chatbot.

A typical user might ask:
💬 "Write me a business proposal for an AI chatbot."

And they’ll get a generic, uninspired response—because the request is too broad.

An analyst, on the other hand, would first ask:

• Who is this proposal for?
• What problem does this chatbot solve?
• What constraints (budget, timeline, tech stack) should be considered?
• What format or structure should the proposal follow?

Then, instead of a vague request, they’d use a structured, high-quality prompt:

💬 "Generate a business proposal for an AI-powered customer service chatbot designed for mid-sized e-commerce companies. The chatbot should reduce response times by 50%, integrate with Shopify and WooCommerce, and fit within a $10,000 implementation budget. The proposal should include a problem statement, solution overview, key features, estimated ROI, and a call to action for decision-makers."

But the process doesn’t stop there.

Just like in a project brainstorming session, the first draft isn’t perfect. You review, refine, and iterate:

• Does the proposal need more detail? Ask AI to expand specific sections.
• Is the tone too formal? Adjust the prompt to request a conversational style.
• Is something missing? Give AI additional context or constraints.

This back-and-forth mirrors how business analysts refine requirements with stakeholders—until they land on a solution that works.

How to Apply the Analyst Mindset to Your AI Prompts

If you want better AI results, start thinking like an analyst:

1. Define the Core Problem Before Asking AI

Before typing anything, ask yourself:
What problem am I trying to solve?

AI works best when given a clear purpose.

2. Identify Key Variables That Shape the Response

An analyst never makes assumptions. Define:

• Audience – Who needs this information?
• Format – List, summary, persuasive argument, structured report?
• Constraints – Budget, timelines, specific requirements?

3. Break It Down Into Steps

Instead of dumping a huge request on AI, build it in layers:

i "List five key content marketing trends for 2025."
ii "Which of these trends are most relevant for B2B SaaS companies?"
iii "Based on these trends, create a 3-month content calendar."

This step-by-step approach ensures accuracy and depth.

4. Brainstorm and Iterate Like You Would with a Team

A great analyst doesn’t walk into a stakeholder meeting and expect perfect answers on the first try. Instead, they:

- Ask clarifying questions

- Reframe based on feedback

- Adjust for missing details

Working with AI is no different. Your first response isn’t the final answer—it’s a starting point. Keep refining, and you’ll land on a solution that works.

Final Thoughts: AI is Easy—If You Know How to Ask the Right Questions

I’ve been diving deep into this process, and I’m excited to share more soon—especially ways to make this even easier to apply in real-world scenarios.

Try it yourself: Next time you use AI, structure your prompt like an analyst. What changes do you notice? Drop a comment—I’d love to hear your experience!