Prompt Playbook: Profitable Problem Discovery PART 3

Remember how in Part 2 I mentioned that once you start looking for business problems, you'll see them everywhere?

Well, there's a flip side to that superpower—it's like opening Pandora's box.

Last month, one of my students texted me in a near-panic. "Kyle, I've created a monster," he said. "I followed your methods and now I've got this firehose of problems—hundreds of them—and I have no idea which ones are worth pursuing. I've made my life harder, not easier!"

I couldn't help but laugh. This is one of those “good problems”!

BUT…it’s still a problem!

When you get really good at spotting business issues that could be solved with AI, you can end up drowning in opportunities. Because, hey, it’s pretty broad!

What she needed—what everyone needs once they've built a substantial Problem Bank—was a system for making sense of it all. A way to categorise, segment, and prioritise these problems to reveal which ones are truly worth pursuing.

And guess what? AI can help us here too.

Let's get started

From Chaos to Clarity: A Simple Filtering Approach

If you've followed Parts 1 and 2, you've likely got dozens (maybe even hundreds) of business problems in your Problem Bank. That's excellent—but it can also be overwhelming.

What you need now is a straightforward filtering approach that helps you identify which problems are worth pursuing. Here's the simple 3-step process we'll use:

1. Basic Categorisation: Organise your problems into meaningful groups

2. Problem Prioritisation: Score problems based on key value dimensions

3. Connection Finding: Identify natural problem clusters or portfolios

This approach is designed to quickly separate the high-potential problems from the rest, without getting bogged down in overly complex analysis.

All sounds very fancy: like I’m dusting off my MBA! But I’ve made a handful of prompts for you to work with - you’re just going to plug in your problems.

Step 1: Basic Categorisation

First, let's organise your Problem Bank into some meaningful categories. The simplest and most useful categorisation is by business function—this naturally aligns with how businesses think about their operations.

Here's a prompt to help you categorise your problems

You are an expert in business problem analysis. I've been collecting business problems that could potentially be solved with AI. Please help me categorise these problems by business function.

For each problem, assign it to one of these business functions:
- Marketing & Sales
- Operations & Logistics
- Finance & Accounting
- Customer Service & Support
- HR & Talent
- Product & Development
- Legal & Compliance
- Other (please specify)

Also, for each problem, identify whether it is primarily:
- An efficiency problem (saving time/effort)
- A quality problem (improving results/reducing errors)
- A revenue problem (increasing sales/opportunities)
- A cost problem (reducing expenses)
- A risk problem (reducing risk/ensuring compliance)

Here's my list of problems:
[Insert your problem list here]

Present your analysis in a simple table with three columns: Problem Description, Business Function, and Problem Type.

After running this prompt with your Problem Bank, you'll have a simple table that organises your problems by function and type. This gives you a basic structure to work with and might already reveal interesting patterns—like a concentration of problems in a particular business function.

This is just the first pass though - we’re going to start cutting it down now.

Step 2: Problem Prioritisation

Now that your problems are organised, it's time to identify which ones have the highest potential value. Not all problems are equally worth solving—some represent much bigger opportunities than others.

The simplest way to prioritise is using a scoring system based on three key dimensions

1. Pain Level: How painful or costly is this problem?

2. Market Size: How many people or companies have this problem?

3. Solution Fit: How well could AI solve this problem?

Here's a prompt to help you prioritise

You are a business opportunity analyst. I need help prioritising business problems to determine which ones represent the best opportunities for AI solutions.

For each problem below, please rate it on a scale of 1-5 (5 being highest) on three dimensions:

1. Pain Level: How painful or costly is this problem for those experiencing it?
2. Market Size: How widespread is this problem across potential customers?
3. Solution Fit: How well-suited is this problem for an AI-based solution?

Then calculate a total score (sum of the three ratings) for each problem.

Problems to evaluate:
[Insert 10-15 problems from your categorised list, focusing on the business functions that had the most problems]

For each problem, provide:
- The three individual scores
- The total score
- A brief explanation of your reasoning
- Any specific aspects that make this problem particularly valuable or challenging

Then rank the problems by total score from highest to lowest.

Now, we are using AI at this stage which is not as good as actually asking customers. BUT we can use this to at least start our winnowing down of problems. I’d recommend using a Research model with this prompt for as much empirical evidence as possible.

Step 3: Finding Connections

The final step is to identify natural connections between your high-scoring problems. We want to start clustering up our problems by similarity. So instead of 10+ problems we might have 2-3 “clusters”.

Once you've identified your top 10-15 problems from the prioritisation step, use this prompt to find connections:

You are a strategic business advisor. I've identified the following high-potential business problems and want to understand how they might naturally connect or relate to each other.

Here are my highest-priority problems:
[Insert your top 10-15 problems from Step 2]

Please help me identify:

1. Natural groupings or clusters of 2-4 problems that are closely related
2. For each cluster, explain why these problems go well together
3. Suggest a simple name or theme for each cluster
4. Recommend which problem in each cluster would be the best one to solve first

Focus on finding meaningful connections where solving one problem would make it easier or more valuable to solve the related problems.

This analysis will reveal natural problem clusters or "portfolios" that could form the foundation of your business offerings. They will be closely associated and complementary. Instead of trying to solve individual problems in isolation, you can develop more comprehensive solutions that address multiple related issues.

We can also (as a bonus) package these up into different product/service levels. Smart!

What's Next?

In Part 4, we'll focus on problem validation. Having identified your most promising problems and clusters, you need to verify they're real, significant, and worth solving before investing in solutions.

We’ve been doing a lot of heavy lifting with AI but the only way to know for sure if these are genuine problems is by taking our solutions to market.

Thankfully we don’t need to fully develop our product or service to do this.

We'll cover initial validation methods, deep AI research techniques, and lean testing approaches to confirm that the problems you've identified are genuinely worth pursuing.

Keep Prompting,

Kyle