From Chaos to Clarity: Problem Solving with AI

🧭 THIS WEEK AT AI SECOND ACT

Howdy, it’s the 10th edition of ASA, and it should be crazy good, fireworks, and generally incredible. Unfortunately, it’s 9.10 pm the night before publication, and I’m just starting the edition. I need a content system asap and to get into a consistent routine. Uggh. Hopefully, though, I can still deliver some value here. This week:

• a possible content shift for AI Second Act - see next code block!

• quick look at Perplexity Labs - incredible!

• using AI to help problem solve

👉 New: I am thinking of structuring a lot more content towards app and tool creation with AI. There are incredible tools like Cursor, Windsurf, Claude Code, and Gemini Code Assist that allow app creation using simple natural language instructions. Is this of interest, high, medium or low? Please hit ‘Reply’ and let me know!

If you’re enjoying this newsletter, please share it! Forward this to a friend/colleague and have them subscribe at aisecondact.com/subscribe. Thanks!

My goal is to make this as valuable and practical as possible as we navigate the new AI era. 🚀

🧰 AI NEWS + LEARNING

Here are a few things I found recently:

• Open AI launches National Academy for AI Instruction - hugely important, just as working with teachers for coding, next is helping skill the next generation of kids with AI skills.

• ‘Vide Coding’ tools I mentioned above - That is, building apps and tools with AI using conversational, natural language prompts - Cursor, Windsurf, Claude Code, Gemini Code Assist. Check out the incredible examples and what you can do with these amazing tools in these links!

• Microsoft AI Health System - Better than doctors at diagnosing complex problems!? With education, I’m hoping AI will hugely impact the health system, which needs help!

Now, to Perplexity Labs! Incredible, amazing, I can not imagine how much time this will save for many different use cases.

• Think of Labs as the Perplexity Deep Research but on a 10x + crazy level. Check out the examples here and a specific F1 report for the research it does, as well as graph/visualization creation. Here’s another example analysing Rivian. Days and weeks of research time become hours with this tool, if not minutes.

🗺️ FEATURED INSIGHT

Work can often be a never-ending tsunami of problem-solving. One after the other. Solve one, relax for 2 seconds, new problem, repeat. Sadly, I don’t recall much in uni on structured problem solving. Well, AI can help.

Ask AI for the Best Problem-Solving Frameworks

Before diving into any messy situation, start by asking AI to recommend the right framework for your specific type of problem.

"I'm facing a [describe problem type - e.g., team conflict, budget overrun, technical decision]. What are the top 3 problem-solving frameworks that would work best for this situation? For each framework, explain when to use it and give me the key steps."

Top Problem-Solving Frameworks (Based on McKinsey/BCG Methods)

1. MECE Framework (Mutually Exclusive, Collectively Exhaustive)

• When to use: Complex business problems with multiple variables

• Best for: Strategic decisions, resource allocation, comprehensive analysis

• Steps: Break the problem into categories that don't overlap but cover everything

• Why it works: Used by McKinsey, BCG, Bain - ensures no gaps or duplication

2. Root Cause Analysis (5 Whys)

• When to use: When symptoms keep recurring or you need to find underlying causes

• Best for: Technical issues, process breakdowns, operational problems

• Steps: Ask "why" 5 times to drill down to the true cause

3. McKinsey 7-Step Problem Solving

• When to use: Major business challenges requiring structured approach

• Best for: Strategic problems, client situations, complex organizational issues

• Steps: Define problem → Structure → Prioritize → Plan → Analyze → Synthesize → Communicate

4. Hypothesis-Driven Problem Solving

• When to use: When data is incomplete or time is limited

• Best for: Fast-moving situations, strategic decisions, consulting projects

• Steps: Form initial hypothesis → Test with data → Adjust → Repeat

5. Issue Trees

• When to use: Breaking down complex problems into manageable parts

• Best for: Team collaboration, assigning workstreams, systematic analysis

• Steps: Create a visual tree showing the problem breakdown in MECE structure

AI-Powered Problem Structuring

Once you know your framework, use AI to structure your messy situation.

Scenario: Project Delay Crisis

The Messy Reality: "Our mobile app release is 3 weeks behind. The iOS team says Android requirements have changed. The Android team says iOS team didn't communicate new API limits. QA found 12 critical bugs yesterday. Marketing already announced the launch date. CEO is asking for daily updates. Budget's getting tight because we're paying contractors overtime."

Prompt for AI:

"Use the Root Cause Analysis framework to help me structure this project delay problem. Here's the situation: [paste messy description above]. 

Please:
1. Identify the main symptom
2. Apply the 5 Whys method
3. Suggest which root cause to tackle first
4. Give me a 30-second explanation I can use with leadership"

AI Output Example:

**Main Symptom:** 3-week project delay

**5 Whys Analysis:**
1. Why is the project delayed? → Critical bugs and platform miscommunication
2. Why did bugs emerge so late? → QA testing started after integration issues
3. Why were there integration issues? → iOS and Android teams worked in silos
4. Why did teams work in silos? → No shared technical specifications document
5. Why no shared specs? → Project kickoff didn't establish cross-platform communication protocols

**Root Cause:** Missing cross-platform communication framework from project start

**Leadership Brief:** "We're delayed because our iOS and Android teams lacked shared technical specs from day one. Fixing the communication process prevents this on future releases. Immediate fix: emergency spec alignment session this week."

Communication Scripts for Different Audiences

Use AI to translate your structured solution into the appropriate language for different stakeholders.

"Based on this problem analysis: [paste your structured solution], help me create 3 different communication versions:

1. Executive summary (30 seconds for leadership)
2. Technical brief (2 minutes for engineering team)  
3. Action items (bullet points for project stakeholders)

Keep each version focused on what that audience needs to know and do."

Pre-Meeting Problem Clarification

Before any problem-solving meeting, use AI to create clarity.

"I'm leading a meeting about [problem description]. There will be [list stakeholder types]. 

Help me prepare:
1. One clear problem statement everyone can agree on
2. Three questions that will focus the discussion
3. Potential solutions to present (not decide, just options)
4. Success criteria for the meeting itself

Keep everything concise - this meeting should solve things, not create more confusion."

Follow-Up and Documentation

After solving the problem, use AI to document lessons learned.

We just resolved [problem description] using [solution implemented]. 

Help me create:
1. A one-page case study for future reference
2. Process improvements to prevent this problem recurring
3. Key lessons for the team
4. Template for similar problems in the future

Focus on what future project managers can learn from this situation."

The Key Insight

AI doesn't solve your problems, it helps you think about them clearly. In tech management, most "complex" problems are simple problems disguised by poor communication and unclear structure.

Use AI as your thinking partner to cut through the noise and find the signal that matters. Remember though, all prompts need iteration and a conversation with AI. Don’t stop at the first response, iterate the prompt and improve the responses.

Future-proof your career with AI
— Brett

👉 Hit “Reply” and share your experience — I read every one!

_{Picture by Jayden Lynch on Unsplash.}