The Leverage Ratio: How to Measure If AI Is Actually Helping You

The Problem With AI Productivity Claims

The AI productivity conversation is broken. People claim 10x productivity gains based on demo videos. Companies report that AI 'saves 5 hours per week' based on surveys where people estimate their own time savings -- which is notoriously unreliable.

The reality is that most people adopt AI tools and then have no idea whether they are actually more productive. They feel faster. They might be faster. But feeling and measuring are different things.

I needed something concrete. Not a survey. Not a feeling. A number I could track daily that would tell me whether my AI investment was paying off or just creating more sophisticated busywork.

The Leverage Ratio: Definition

The Leverage Ratio is simple:

Leverage Ratio = Total Output Value / Gregor Hours Invested

The scale:

1x -- You did it manually. No leverage. You spent 4 hours and got 4 hours of output.

2x -- AI did 50% of the work. You spent 2 hours and got 4 hours of output.

3-5x -- AI does most of the execution. You direct and review. This is the target zone.

5x+ -- The system runs without you for extended periods. This is the aspiration.

The key insight: this is not about time tracking. It is about asking one question at the end of every day -- did I do this, or did I direct something else to do this?

How I Track It

I maintain a simple weekly log with four columns:

1. Delegation rate: What percentage of tasks this week were executed by agents vs. by me?

2. Bottleneck count: How many times did work stall because it was waiting for my input?

3. Pipeline per Gregor-hour: How much active pipeline value exists relative to the hours I spent on sales this week?

4. Output inventory: What did agents produce this week that I would have otherwise done myself?

The tracking is not precise. It does not need to be. The value is in the trend -- is the ratio going up, staying flat, or going down? A flat ratio means AI adoption is not actually reducing your workload. A declining ratio means you are spending more time managing AI than it is saving you.

My Actual Numbers

When I started tracking in early 2026, my Leverage Ratio was approximately 1.5x. I was using AI for drafting and research, but I was still doing most of the execution myself. The agents existed but I was not trusting them with real work.

As of April 2026, the ratio is approaching 2.5x. The W15 weekly plan shows 'Total Gregor time: ~1.5 hours. Everything else is agent-owned.' That is a real data point, not a projection.

The target is 3-5x by end of Q2 2026. That means spending 10-15 hours per week on direction and review, while agents produce the equivalent of 40-60 hours of work.

Honest caveat: the ratio fluctuates. Weeks with client discovery calls (which require me personally) have lower ratios. Weeks focused on content and outbound (which agents handle well) have higher ratios. The metric is most useful as a weekly average, not a daily number.

The Should Gregor Be Doing This Filter

The Leverage Ratio led to a practical decision filter I apply to every task:

Level 1: Can an AI agent do this right now? Delegate immediately. Do not touch it.

Level 2: Can an AI agent do 80% of this? AI drafts, I review. This is 2-3x leverage on that task.

Level 3: Does this require my relationships or judgment? I do it, but AI prepares everything -- research, talking points, background, draft materials.

Level 4: Am I doing this out of habit, not necessity? Stop. Find the alternative. This is where the biggest leverage gains hide.

Most people skip Level 4. They keep doing things because they have always done them, not because they are the best use of their time. The Leverage Ratio makes Level 4 visible -- if you are consistently at 1.5x despite having AI tools, the problem is probably habit, not tooling.

Common Mistakes When Measuring AI Leverage

Mistake 1: Counting time saved instead of output gained. 'AI saved me 2 hours' is meaningless if you spent those 2 hours on something equally low-leverage. The ratio measures output value, not time savings.

Mistake 2: Measuring individual tasks instead of the system. AI might be fast at drafting an email but if you spend 30 minutes reviewing it, the net leverage on that task is low. Measure the end-to-end workflow, not the AI step in isolation.

Mistake 3: Ignoring the setup cost. Building an AI agent takes time. The Leverage Ratio will be below 1x during setup. That is expected. The question is whether it crosses 2x within 2-4 weeks and keeps climbing.

Mistake 4: Optimizing for the wrong metric. Some tasks should not be delegated to AI even if they could be. Client relationship building, strategic decisions, and anything requiring genuine empathy -- these are not leverage targets. They are your competitive advantage.

How to Start Tracking Your Own Leverage Ratio

You do not need a fancy tool. A weekly note with three questions:

1. What did I produce this week? List everything: emails sent, content published, proposals delivered, code shipped, decisions made.

2. For each item: did I do this or direct something to do this? Mark each as 'me' or 'delegated' (to AI or human).

3. How many hours did I spend on direction and review vs. execution?

The ratio is direction+review hours divided into total output value. Start tracking. Do it for 4 weeks. The trend will tell you whether AI is actually helping or whether you are just a more sophisticated typist.