Why Yearly Goals Fail—And the AI System I Built to Fix It

Every January, I set new year goals.

“This year I’ll exercise. Write more. Build my product.”

And then?

February: Still remember. Still doing it.

March: Busy. I’ll get back to it.

June: “Wait, what goals did I set again?”

December: “Forget it. Next year.”

I’ve played this script for years.

The Method: 12 Week Year

Last year, I tried something different.

The core concept is simple: a year is too long. When you have 12 months, your brain switches to power-saving mode. There’s always time. Tomorrow.

So instead, treat 12 weeks as a full year.

12 weeks is short enough to create urgency. Long enough to accomplish something meaningful. And you get 4 chances to start over each year.

The framework comes with an execution system: Lead vs Lag indicators, weekly scoring, WAM accountability meetings. I wrote a complete guide—I won’t repeat it here.

The point is: I started using it. Set goals, broke them into weekly tasks, scored each week.

But I quickly found a problem.

The Problem: Frameworks Aren’t Enough

I still forgot to check.

OKR, GTD, 12 Week Year—I’ve tried them all. They share the same problem: exciting to set up, frustrating to execute.

Because you have to remember to check. Remember to update. Remember to track.

“Remembering” itself is a burden.

How much did I spend this month? How many goals did I complete last week? That flag I set three months ago—where am I now?

I don’t know.

Data scattered across apps. To-dos in six different places, none of them complete.

The problem isn’t willpower. The problem is not being able to see yourself.

Can’t see where you are, where you’re going, how far you have left.

The Solution: Using AI to Track My Goals

So I used Claude Code to build my own system.

This system doesn’t help me “set goals”—I still have to figure those out myself.

It helps me “track goals.”

I call it NorthStar—one glance tells me where I am and where to go.

My Personal AI System

The system works in two scenarios:

At my computer (deep work): Data lives in a local Cockpit folder, all Markdown files. Every morning I plan in the terminal with Claude Code—it reads and writes these files directly, updating progress, reviewing yesterday, planning today.

Away from computer (fragmented time): LINE pushes the Daily Brief, and I can toss ideas into it anytime. Weekly tasks sync to Google Tasks—visible and checkable on my phone. Calendar pulls from Google Calendar.

One source of truth, stored in Cockpit. Two scenarios, each taking what it needs.

Feature 1: Daily Brief

Every day at 5 AM, the system runs automatically:

Pulls this week’s Work Items and progress from Google Tasks
Scans emails, flags important ones
Fetches trending content from Reddit, HN, X, filtered by topics I care about
Compiles everything into a Daily Brief

At 6:47 AM, my phone buzzes.

LINE notification:

☀️ Today’s Brief is ready.

📊 This week’s goal completion: 73%, 2 Work Items remaining.

📧 3 important emails to handle.

💡 There’s a Reddit discussion worth reading.

Before I even open my eyes, everything I need to know is already organized.

Feature 2: Goal Tracking

This system uses the 12 Week Year framework mentioned earlier.

Core concept: distinguish between Lead (controllable actions) and Lag (desired results).

Before, I only watched Lag—“Revenue needs to grow 20% this month!” Then I’d check the numbers daily, anxious about why they weren’t moving.

Now I watch Lead—“This week: publish 2 articles, contact 5 clients.” Done? Results follow. Not done? At least I know where the problem is.

The system tells me daily: “5 Work Items this week, 3 done, 2 remaining.” No guessing, no trying to remember.

Once a week, WAM (Weekly Accountability Meeting): What was last week’s completion rate? What’s planned for this week? If Lead actions are done but Lag isn’t moving—is it the wrong actions, or the wrong strategy?

Not just tracking progress. Strategic calibration.

Feature 3: Content Discovery

I write. The hardest part isn’t writing—it’s not knowing what to write.

Before, I’d spend 1-2 hours scrolling Reddit, Hacker News, X, looking for interesting discussions. But I’d scroll and forget why. By the 30th post, I’d forgotten what the 1st one said.

Time spent, inspiration missed.

Now the system runs automatically each morning:

Fetches trending content
AI filters for high-engagement posts relevant to me
Generates summaries, 1-2 sentences each
Categorizes: better for FB or Blog

When I wake up, inspiration is already in the Brief:

💡 Today’s content inspiration:

Reddit: Solo founder hit $10K MRR → good for FB

@levelsio: AI coding observations → write your take

No more searching for needles in haystacks.

Feature 4: Quick Inbox Capture

In the shower, a great idea hits. “Remember to write this down when I get out!”

Then I get out and forget.

Now whenever something comes to mind, I toss it into LINE.

“Add this feature”—toss it in.

“Write about this topic”—toss it in.

No categorizing, no organizing. The system automatically saves it to Cockpit’s inbox folder.

Every morning, the Brief reminds me: ”📥 Inbox has 5 items to process.”

Open CLI, Claude asks: “Add this idea to backlog? Or delete?”

Valuable ones stay, useless ones go. Brain can clear out.

Feature 5: Email Management

My inbox used to look like this:

147 unread. Maybe 3 important. But finding them took 40 minutes.

Newsletters, ads, notifications, spam—the truly important emails buried underneath.

Now the system scans, categorizes, and drafts replies automatically each morning:

3 important emails today:

Client A asking for quote — suggest replying today

Partner B confirming meeting time

Bank notice: credit card statement

Drafts are already sitting in my Mail app’s drafts folder. I open email, glance, change two words, hit send.

Before: 40 minutes. Now: 10 minutes.

Feature 6: Visual Dashboard

Numbers need to be understood at a glance.

I built my own web interface: dashboard, calendar, to-do list. Not Excel, not Word, not scattered Markdown files. Built from scratch based on my own needs and habits.

Weekly progress, completion rate, goal status—one screen shows everything.

This system isn’t perfect yet. Using it daily, hitting issues daily, fixing daily.

But perfection isn’t the point.

The point is: it’s a system that can be tested and corrected.

Before, I worked by feel—didn’t even know where I went wrong. Now there are numbers, records. When wrong, I can look back. Next time, I can adjust.

Work Items completion at 70% this week? Why? Set too many goals? Got interrupted? When you can see it, you can fix it.

Before, I spent an hour each morning “getting ready to work”—checking email, organizing to-dos, scrolling social media, trying to remember what I did yesterday.

Now I open my phone, the Brief is there. Then I start working.

Seeing Yourself

Yearly goals fail year after year, not because of laziness.

It’s because you can’t see yourself.

Can’t see where you are, where you’re going, how far you have left.

Thanks to AI, I can build a system that’s truly mine. Not an app I bought, not someone else’s template—built piece by piece based on my own needs, my own habits.

When you can finally see where you stand—

I think moving forward isn’t so hard after all.

This is the third article in the “Building a Personal AI System” series:

From Obsidian to Claude Code: Building a Programmable Life — The starting point, three core principles
The Complete 12 Week Year Guide — Lead vs Lag, scoring, WAM accountability
This article — 12 Week Year + AI integration in practice

Want more content like this? Subscribe to my newsletter