The Question you need to ask yourself: How many AI tools are you paying for that have become nothing more than expensive bookmarks? You know the ones—they wooed you with a slick demo, promised to change your life, and now just sit there collecting dust like a Tinder match that never texted back.

Firstly, you should: Upload a document listing every AI tool and subscription you are currently paying for.

Query your LLM with this prompt:

Act as a ruthless efficiency consultant with zero patience for sunk cost fallacy. For each tool, interrogate me:

1. When did I last use this for actual output—not "exploring," not "testing," not "I opened it once"?
2. What specific, measurable result did it produce? Give me KPIs: How many finished pieces of work came from this tool in the last 30 days? How many times did it directly contribute to something that shipped, published, or generated revenue?
3. Is that usage one-time or recurring? A tool that helped once six months ago is not earning its subscription.
4. If I cancelled it tomorrow, what would actually break?

Be merciless. "I might need it someday" is cope, not strategy. "It's only $20/month" is how $240/year disappears. "I haven't learned it properly yet" means I never will.

After the audit, sort my tools into three buckets:

**Keep** (must meet ALL criteria):

• Used 4+ times in the last 30 days with measurable output each time
• Can directly attribute at least one shipped deliverable, published piece, or revenue-generating outcome per month
• If it disappeared tomorrow, I'd repurchase within a week because nothing else does what it does
• Cost per use is under $1

**Kill** (meets ANY of these):

• Zero uses in the last 30 days
• Cannot name a single concrete output from the last 90 days
• If it disappeared, I wouldn't notice for a month—or ever
• Cost per use is over $10, or infinite because I didn't use it at all