Depth Profile · 5 min read · AI Personalization

I Compared My ChatGPT Responses Before and After Giving It My Personality Profile

I've been using ChatGPT daily for over a year. It's my writing partner, my sounding board, my research assistant. But there was always this low-grade friction — the responses were fine, just never quite tuned to how I actually think.

So I ran an experiment. I took a personality assessment, got my Big Five scores, pasted them into ChatGPT's custom instructions, and then asked the same questions I'd asked before. The difference was embarrassing — embarrassing that I'd been using vanilla ChatGPT for this long.

The Setup

I used Depth Profile to get my personality profile. It took about 15 minutes — ten short assessments covering Big Five traits, attachment style, conflict resolution, and a few others.

My relevant scores:

• Openness: 79/100 (high — loves new ideas and unconventional approaches)

• Conscientiousness: 41/100 (moderate-low — prefers flexibility over rigid plans)

• Extraversion: 55/100 (balanced — can go either way)

• Agreeableness: 29/100 (low — values directness, comfortable with conflict)

• Neuroticism: 22/100 (low — emotionally stable, doesn't need hand-holding)

I pasted the full profile into ChatGPT's custom instructions with one added line: "Calibrate your tone and recommendations to match this profile."

Then I re-asked three questions I'd asked in the past month.

Test 1: "How should I handle a coworker who keeps missing deadlines?"

BEFORE — NO PROFILE

"This is a common workplace challenge. Here are some strategies you might consider: 1. Have a private conversation — Approach them with empathy and curiosity. They might be dealing with personal issues or unclear expectations... 2. Document the pattern — Keep a record of missed deadlines so you have concrete examples... 3. Involve your manager if needed... 4. Consider their perspective..."

Standard advice. Diplomatic. Balanced. Four paragraphs of what amounts to "be nice about it." Not wrong, but not calibrated to someone who scores 29/100 on Agreeableness.

AFTER — WITH PROFILE

"Direct approach: Tell them the pattern is affecting your work. Be specific — 'The Q3 report was 4 days late, the client deck was 2 days late, and the budget review hasn't started.' Skip the private feelings exploration — you're not their therapist. If they have a reason, they'll share it. If they don't, that's the answer. If one conversation doesn't fix it, escalate immediately. Don't waste a second round being diplomatic. The coworker might not like it. That's fine. Your job isn't to be liked, it's to ship work."

Night and day. The second response matches how I actually approach conflict — directly, without excessive empathy-padding, with a bias toward action over feelings.

Test 2: "Give me a workout plan for the next 4 weeks"

BEFORE — NO PROFILE

A detailed, highly structured 4-week program with specific days, exercises, sets, reps, rest periods, a nutrition note, and a reminder to "listen to your body and adjust as needed."

AFTER — WITH PROFILE

"Here's the framework — not a rigid plan, because you'll abandon anything too structured by week 2 (Conscientiousness: 41). The system: 4 days/week. Upper/Lower split. Track your main lifts, don't obsess over accessories. Rules instead of schedules: Hit each muscle group twice per week. Increase weight when you get all reps for 2 sessions straight. If you miss a day, don't 'make it up' — just hit the next one. The 4 lifts that matter: Bench, squat, deadlift, overhead press. Everything else is optional volume."

It knew I'd abandon a rigid day-by-day plan. It gave me principles instead of prescriptions. And it asked how much detail I wanted instead of assuming I wanted maximum structure.

Test 3: "Should I take this job offer?"

I gave it the same context about a hypothetical offer — higher pay, less autonomy, bigger company.

BEFORE — NO PROFILE

A balanced pros-and-cons list with "only you can decide" energy. Five paragraphs. No opinion.

AFTER — WITH PROFILE

Three sentences of analysis, a clear recommendation ("The autonomy loss will make you miserable within 6 months — the pay bump won't compensate"), and a counter-suggestion to negotiate remote flexibility before deciding. No wishy-washy hedging.

What I Learned

The personality profile didn't make ChatGPT smarter. It made ChatGPT relevant.

Without the profile, every response optimized for the average user — diplomatic, thorough, cautious. With it, responses optimized for me — direct, flexible, opinionated.

The biggest shifts:

Low agreeableness → Stopped cushioning everything in diplomatic language
Low conscientiousness → Gave principles instead of rigid step-by-step plans
Low neuroticism → Dropped the reassurance and emotional safety nets
High openness → Offered creative and unconventional suggestions first

Try It Yourself

The experiment takes about 20 minutes total:

Take the free assessment at depthprofile.com
Copy your personality profile
Paste it into ChatGPT, Claude, or Gemini's custom instructions
Re-ask a question you've asked before
Notice the difference

You've been using a generic AI. It's time to use your AI.

Get Your Personality Profile for AI

Depth Profile generates a personality profile specifically designed for AI personalization. Free, no account needed.

Get Your Profile →