You've tried building Excel models with ChatGPT or Claude. Sometimes it works perfectly—you describe what you need, and minutes later you have a working DCF. Other times, you iterate ten times and still get something that's almost right but not quite.
What's the difference?
Most people think it's about better prompts. Write clearer instructions, add more detail, and the AI will understand.
But that's not it.
The Real Problem: The Context Gap
Here's what's actually happening:
When you ask an AI to build a model, your head is full of context. You know:
- The deal structure
- Your firm's modeling conventions
- What "cash flow" means in this situation
- How you structured similar models before
- Your formatting preferences
The AI knows none of this.
Every conversation starts from zero. It doesn't remember your last model. It doesn't know your preferences. It has no context about your firm, your deal, or how you like to work.
This gap between what's in your head and what the AI sees is the source of most frustration.
You say "build me a real estate model" and assume the AI understands what that means to you. But without context, it has to guess at structure, time periods, outputs, formatting—everything.
The Skill Nobody Talks About
The analysts who get consistently good results from AI aren't the ones with the best prompts.
They're the ones who've mastered context management.
They know:
- What context actually matters (and what's just noise)
- How to describe existing models efficiently
- When to collaborate vs. when to take control
- Which mode of interaction fits each task
It's a meta-skill that sits underneath all the tactical advice about prompting.
The Framework
In our complete guide, we break down context management into four core areas:
1. The Context Gap Understanding what the AI actually needs to know—and what's just noise that clouds the conversation.
2. The Four Collaboration Modes When to use generative mode vs. advisory mode vs. pair building vs. debugging. Each has its place, and choosing wrong makes everything harder.
3. Describing What You Have How to communicate about existing models efficiently, whether you're modifying a formula or restructuring an entire sheet.
4. Practical Patterns Specific approaches that work reliably: skeleton-first building, example-based prompting, constraint-based specifications, and knowing when to stop iterating.
Why This Matters Now
AI tools for Excel are getting better fast. But the gap between people who struggle with them and people who use them effortlessly is widening.
It's not about technical skill. It's not about Excel knowledge.
It's about understanding how to collaborate with AI effectively.
Master context management, and everything else gets easier. You iterate less. You get better outputs. You work faster. You can tackle more complex models.
The investment in learning this skill pays off in every session.
Learn the Framework
We've put together a complete guide on context management for building Excel models with AI.
It covers:
- The exact context checklist to use before every session
- When to switch between the four collaboration modes
- Five proven prompt patterns that work reliably
- How to know when to iterate vs. when to take control
- Building your practice over time
