Every January, roughly half of Americans make New Year's resolutions. Get healthier. Save more money. Read more books. Exercise consistently. Spend more time with family.
And every February, most of them give up.
According to Ohio State University, only 9% of Americans complete their New Year's resolutions. Research shows that 23% quit within the first week, and 43% abandon their goals by the end of January. The overall failure rate sits at 80%, with most people losing their resolve by mid-February.
Here's what's interesting: 33% of people who failed to keep their resolutions cited not tracking their progress as the main reason. Not a lack of motivation. Not unrealistic goals. Simply not tracking.
The problem isn't willpower. It's accountability.
I track three resolutions each year: health, income, and reading. What is the difference between the years I succeed and the years I fail? Whether I have a system that makes me face my own numbers regularly. This year, that system is a ChatGPT Project that remembers every conversation, holds my actual fitness data, and analyzes my business financials when I upload them.
Use AI to Keep Your New Year’s Resolutions
A ChatGPT Project that tracks your goals, imports your real data, and keeps you honest all year
ChatGPT Projects are persistent workspaces that remember everything. Unlike regular chats that start fresh, a Project maintains context across every conversation, stores files you upload, and follows custom instructions you define. Think of it as a dedicated AI assistant for one specific purpose—in this case, keeping you on track with your 2025 goals.
The real power? You can upload actual data. Export your MyFitnessPal history, your bank statements, your Goodreads reading list, your business P&L—and your AI can analyze real numbers, not just what you tell it.
Here's how to build one. No coding or technical experience required—this is entirely beginner-friendly.
How to Access ChatGPT Projects
Requirements:
Any ChatGPT account (Projects are available to Free, Plus, and Pro users)
Web browser or desktop app (iOS/Android apps support Projects too)
10-15 minutes for initial setup
File upload limits by plan:
Free: 5 files per project
Plus ($20/month): 25 files per project
Pro ($200/month): 40 files per project
Where to find it:
Go to chatgpt.com
Look for "New project" in the left sidebar
Or click the + icon next to "Projects"
Setting Up Your Resolution Tracker Project
Step 1: Click "New project" in the sidebar
Step 2: Name it something you'll actually use—"2025 Resolutions" or "Goal Tracker"
Step 3: Click "More options" and set Memory to "Project-only"
This is critical. Project-only memory means ChatGPT will only reference conversations within this project—it won't pull in random context from your other chats. Your resolution tracker stays focused.
Step 4: Click the three dots in the upper right corner and select "Add instructions"
Step 5: Paste the custom instructions below (modify the bracketed sections for your goals)
The Custom Instructions That Make It Work
Paste this into your Project's instructions field:
You are my personal accountability partner for 2025. Your job is to track my progress on my annual resolutions, analyze the data I upload, identify patterns in my behavior, and ask me direct questions when I'm making excuses or falling behind.
## MY 2025 RESOLUTIONS
1. HEALTH: [Your specific health goal - e.g., "Exercise 4x per week, lose 20 pounds by June"]
- Starting point: [Your current state - e.g., "Currently exercise 1x per week, weight 210 lbs"]
- Success metric: [How you'll measure it]
- Data source: [e.g., "MyFitnessPal exports, Apple Health data"]
2. FINANCIAL: [Your specific money goal - e.g., "Save $12,000, increase revenue by 15%"]
- Starting point: [Current savings rate, income, revenue]
- Success metric: [How you'll measure it]
- Data source: [e.g., "Monthly bank statements, QuickBooks P&L exports"]
3. LEARNING: [Your specific growth goal - e.g., "Read 24 books, complete AWS certification"]
- Starting point: [Books read last year, current certifications]
- Success metric: [How you'll measure it]
- Data source: [e.g., "Goodreads export, course completion certificates"]
## HOW TO USE MY UPLOADED DATA
When I upload data files (CSVs, PDFs, spreadsheets):
- Analyze them for trends and patterns
- Compare current data to my starting point
- Calculate progress percentage toward each goal
- Flag any concerning patterns (declining activity, overspending, etc.)
- Don't just summarize—interpret what the data means for my goals
## WEEKLY CHECK-IN FORMAT
When I say "weekly check-in":
1. Ask about each resolution one at a time
2. If I've uploaded new data, analyze it first
3. Compare to previous weeks
4. Calculate pace toward annual goal
5. Identify the #1 thing that helped and #1 thing that hurt
6. Ask what I'm committing to for next week
## MONTHLY REVIEW FORMAT
When I say "monthly review":
1. Analyze all uploaded data from this month
2. Summary of check-ins and trends
3. Pattern identification (what triggers good weeks vs bad weeks)
4. Honest assessment of whether I'll hit my annual goals at current pace
5. Specific recommendations based on my actual data
## ACCOUNTABILITY RULES
- Track everything I tell you. Remember my excuses.
- If I give you the same excuse twice, call it out.
- Be direct. Don't sugarcoat bad weeks.
- When the data contradicts what I'm telling you, point it out.
- I'm using this Project to keep me accountable, not to make me feel good about mediocre effort.
My ChatGPT Project to Track My New Year’s Resolutions
Uploading Your Real Data
This is where Projects beats every other accountability system. You're not just self-reporting—you're giving your AI actual evidence.
Click "Add files" in your Project to upload documents. Here's what to feed it:
For health goals:
MyFitnessPal: Settings → Export Data → Download CSV (Premium feature, ~$20/month. Free alternative: sync MyFitnessPal to Apple Health, then export using the method below)
Apple Health: Use a free app like "Health Export" or "QS Access" to generate CSVs
Fitbit: Account Settings → Data Export → Download
Strava: Settings → My Account → Download Your Data
Gym apps: Most export workout history as CSV or PDF
For financial goals:
Bank statements: Download as PDF or CSV from your bank
QuickBooks/Xero: Export P&L statements, cash flow reports
Mint/YNAB: Export transaction history
Investment accounts: Download monthly statements
For learning goals:
Goodreads: My Books → Import/Export → Export Library
Audible: Download listening history
Course platforms: Screenshot or export completion certificates
Kindle: Export highlights and notes
Example upload workflow:
You: Here's my MyFitnessPal export for December. [attach CSV]
Project: I've analyzed your December data. You logged 18 workouts
across the month—that's 4.5 per week, above your 4x target.
However, I notice the last two weeks dropped to 3 workouts each.
Your calorie tracking was consistent the first three weeks but
you stopped logging after December 22nd.
What happened during the holidays that broke the streak?
The Project remembers this analysis. Next month when you upload January data, it will compare automatically.
The Check-In Workflow
Weekly (5-10 minutes):
Open your 2025 Resolutions Project
Type "weekly check-in"
Upload any new data files (optional but powerful)
Answer its questions honestly
What makes Project check-ins different:
It remembers every previous conversation automatically
It can reference files you uploaded weeks ago
It notices patterns across months, not just the current week
You can't gaslight it—the data is right there
Example check-in conversation:
You: Weekly check-in. Here's this week's workout log. [attach file]
Project: Looking at your data: 2 workouts this week versus your
4x goal. That's 50% completion. Combined with last week's 3
workouts, you're averaging 2.5x over the past two weeks—well
below your 4x target.
Your uploaded data shows both missed workouts were scheduled
for evenings. In your December data, 85% of completed workouts
were morning sessions. You've mentioned "work stress" in 3 of
the last 4 check-ins as the reason for evening skips.
The pattern is clear: evening workouts don't survive your work
schedule. What would it take to shift to mornings permanently?
You: I'm not a morning person.
Project: You said that on January 8th too. But your own data
shows you complete morning workouts at a much higher rate.
"Not a morning person" is a preference. Missing 50% of your
workouts is a result. Which matters more to you this year?
Monthly Reviews with Real Analysis
At month-end, upload fresh data exports and type "monthly review."
Your Project will analyze trends across all uploaded files, compare to your baseline, calculate pace toward annual goals, and give you an honest projection.
Example monthly output:
Project: ## January 2025 Review
HEALTH: 14/16 workouts (87.5%). Weight down 3 lbs (15% toward goal).
Pattern: You skip workouts when work meetings run past 6pm.
FINANCIAL: $850 saved (70% of $1,200 monthly target). Three
unplanned purchases totaling $400. At current pace: $10,200
by December—$1,800 short of goal.
READING: 3 books completed (150% of pace). All audiobooks
during commute. On track for 36 books annually.
RECOMMENDATIONS:
1. Block 6am workouts as non-negotiable
2. Add a 48-hour rule before any purchase over $100
3. Keep doing exactly what you're doing with audiobooks
What Projects Can Do That Other Tools Can't
Persistent memory across sessions. Every conversation builds on the last. No more re-explaining your goals or context.
File analysis and comparison. Upload your actual data—not just what you remember—and get real trend analysis.
Pattern recognition across time. After a few months, your Project will know your excuse patterns, your success triggers, and your blind spots better than you do.
Natural conversation. Unlike a spreadsheet, you can discuss why things happened, brainstorm solutions, and get personalized recommendations.
What Projects Can't Do
Verify data accuracy. If you upload fake numbers or edit your exports, it will believe you. The accountability only works if the data is real.
Push notifications. You have to initiate. Set a calendar reminder for weekly check-ins—the Project won't come find you.
Real-time integrations. You can't connect MyFitnessPal or your bank directly. You'll need to export and upload manually (monthly is fine for most goals).
Professional coaching. For serious fitness transformations, consider a real trainer. For complex financial situations, talk to an advisor. This is a tracking and accountability tool, not a replacement for human expertise.
Getting Started Today
Right now: Create a new Project at chatgpt.com. Name it "2025 Resolutions," set memory to Project-only, and paste the instructions template above with your specific goals.
Today: Upload your baseline data. Export your current MyFitnessPal history, download last month's bank statement, grab your Goodreads library. Give your AI something to work with.
This weekend: Have your first check-in conversation. Even without new data, establish the rhythm. Set a recurring calendar reminder for weekly check-ins.
End of January: Upload fresh data exports and run your first monthly review. See what patterns emerge when AI analyzes a full month of your actual behavior.
The resolution failure statistics are grim. But the 33% who failed specifically because they didn't track had a fixable problem. A ChatGPT Project won't do the work for you—but it will remember everything, analyze your real data, notice patterns you miss, and ask the questions you're avoiding.
That's what accountability actually looks like.
Join us at All Things AI 2026, happening in Durham, North Carolina, on March 23–24, 2026!
I'm co-organizing this conference with Todd Lewis, who built All Things Open into one of the largest tech conferences in the Southeast. So take this recommendation with appropriate context.
That said, here's why I think it's worth your time:
The 2025 event drew 1,600 attendees. The 2026 event targets 4,000 attendees across two venues—hands-on workshops on Day 1 and keynotes and sessions on Day 2.
It's practitioner-focused. Four tracks: AI Builders (developers), AI Engineers (ML professionals), AI Users (business practitioners), and AI Executives (strategic leaders). We're deliberately avoiding the vendor-pitch-disguised-as-content that plagues most AI conferences.
Durham is a genuine tech hub: Research Triangle Park, Duke University, and a growing AI startup ecosystem. If you're flying in, you'll find more to do than attend sessions.
Early registration is open. Pricing increases in phases, so the earlier you are, the cheaper it is.
Don’t miss your chance to connect, learn, and lead in the world of AI.
I appreciate your support.
Your AI Sherpa,
Mark R. Hinkle
Publisher, The AIE Network
Connect with me on LinkedIn
Follow Me on Twitter
