Most PM knowledge bases fail the same way: the content is solid, but the path to it is too long. Every extra click between your user and your best thinking is a percentage of engagement you will never recover. If the value exists but the access is painful, it effectively does not exist.

I launched The Vault with 100+ prompts in April. Downloads on launch week were strong. Return visits were not. By week three I pulled my analytics and traced the exact journey a subscriber would take from inbox to actual prompt: open the email, click the website link, log in to the account, navigate to the right category, scroll to find the prompt, copy it, switch to their AI tool.

Seven steps between the value I built and the moment they could actually use it.

The distance between completing the action was huge.

I had been adding new prompts every week. Stronger titles, cleaner category labels, better descriptions. None of it moved the return engagement because I was improving the wrong layer. The friction was not inside the library. It was in the path to get there.

The moment I counted the steps out loud, the fix became obvious. But I would not have counted them if I had not run this audit.

The Prompt Logic

This prompt works because it forces you to count the literal physical steps, not the conceptual journey. Most PMs think about UX in terms of flows and funnels. This audit makes you count tabs.

The “Intent-Unit” concept is the mechanism that makes it stick. Instead of asking “what does my user want from this product overall,” you ask: what is the single atomic thing they came to get right now? Once you name that unit, the distance between your user and that thing becomes both obvious and embarrassing.

The vertical slice in step 3 is where the real work happens. It forces the question: can I deliver this Intent-Unit directly into the tool my user already has open? For me, the answer was yes — but only if I stopped treating my website as the delivery layer.

I built a Claude-native MCP that brings Vault prompts directly into the subscriber’s AI session. The website became optional. The friction dropped to one step.

What did I learn? 🤔

“Information Distance” is the number of actions between a user’s current context and your product’s core value. In 2026, any distance greater than two steps is churn waiting to happen. Your product is not your content. It is the interface to that content.

The Prompt

Role: Product Architect.
Context: I have a product with valuable logic (e.g., templates, blueprints, data) but user engagement is low. I suspect the friction of accessing it is too high.

Task:
1. List the exact, physical steps a user must take to access and use my product's core logic, starting from their primary work environment (e.g., Claude, Slack, VS Code).
2. Isolate the single most valuable "Intent-Unit" a user seeks.
3. Propose a "Vertical Slice" to deliver this Intent-Unit directly into the user's primary tool to eliminate workflow friction.

Input: [Describe your product, its core logic, and the user's primary work tool].

How to Run It in Claude or Cursor

Open a new Claude session and paste the full prompt. Fill in the Input field with an honest description of your product and what your users actually do when they arrive. Be specific about the work tool — “email inbox” is a completely different problem than “Claude” or “VS Code.”

Before you run it, write down your current step count. Just count them by hand, starting from wherever your user is when they decide they want the thing you built. Then run the audit and read the proposed vertical slice.

If step 1 returns more than four actions before the user gets value, you have an architecture problem to solve before you write any more content.

Run the prompt again after you ship the fix. The 𝚫 delta between the two step counts is your friction score. That number is more useful than any conversion rate you will find in a dashboard.

Why This Is a Competitive Edge

Most PMs audit the product itself. They add features, refine onboarding, rewrite copy. This prompt flips the direction and audits the gap between the product and the user’s existing workflow.

The products that win in 2026 are not the most feature-rich. They are the ones with the shortest distance between the user’s intent and the delivered value. Running this audit on your own product — quarterly, without being asked — is how you catch the abandonment signal before it shows up in your churn numbers.

The subscribers you lose to friction rarely complain. They just stop coming back.