The result
I thought I needed to build a new home for practical AI content. The inventory showed that Building Empires was already the closest fit. The right move was to revive and refocus it, while keeping Atla and Habits That Matter separate.
Before you begin
Work on a copy or use read-only instructions first. Do not give an AI access to passwords, tax records, medical records, private client data, API keys, or environment files. The point is to understand the work — not hand over every private file you own.
Step 1: Inventory
Start with observation only. Make the tool tell you what exists and what state it is in.
Search this folder and build an inventory of the work that is already here. For each relevant item, tell me: 1. what it is; 2. the audience and outcome it appears to serve; 3. whether it is live, usable, unfinished, duplicated, or retired; 4. what other files or projects it connects to. Do not move, delete, publish, or edit anything. Exclude secrets, environment files, personal records, and unrelated code.
Step 2: Cluster by outcome
Folder names usually reflect when something was created. Outcomes reveal what can be combined without creating a confusing business.
Group the usable work by the outcome it creates — not by file type or folder name. For each group, identify: - the strongest existing brand or container; - the working assets already present; - the missing pieces required to put it back to work this week; - anything that should stay separate. Flag contradictions and stale claims. Do not invent a new brand.
Step 3: Make one decision
Do not ask for fifty ideas. Give the AI your current focus and force it to choose the nearest existing asset with evidence.
My focus is: [describe the one outcome you want now]. Choose the existing asset that is closest to that outcome. Recommend the smallest version I can finish this week using what already exists. Return: 1. keep; 2. change; 3. ignore for now; 4. the first three actions; 5. the evidence you used to make the choice.
Where AI failed
- It could identify close matches, but the final brand decision still required human judgment.
- Old plans contradicted newer priorities, so dates and current intent mattered.
- Technical reuse did not automatically mean strategic fit. Atla could share infrastructure, but it should remain its own product.
The verdict
DONE. This is worth doing before you name, buy, or build another property. AI is excellent at recovering forgotten work when the task is read-only, the outcome is clear, and the person makes the final call.