Vibe Intensity Profile: Stories Become Numbers

Traditional research takes weeks and costs thousands. What if you could get strategic answers in an hour?

VIP (Vibe Intensity Profiles) makes this possible. Stories become numbers. Nuance becomes visible.

Explore distributions, ratings, and trends:

• Which marketing campaign will drive results? Test three concepts. Launch the winner.
• How will users respond to this design? Explore reactions across segments before building.
• Where do we sit in the competitive landscape? Map perceptions fast.
• Is this research brief worth £50K? Simulate first. Commission smarter.

VIP

Standard surveys force a choice: Pick one - Strongly Disagree, Disagree, Neutral, Agree, Strongly Agree.

Real opinions aren't single points. "I mostly agree, with some reservations."

VIP asks a question and catches the response's whole spectrum: 60% Agree, 30% Strongly Agree, 10% Neutral.

One response. Full range of feeling.

Note: I Made It Up includes 6 scale types (Purchase Intent, Importance, Satisfaction, Likelihood, Agreement, and Sentiment). We can create custom scales on request for specific research needs.

How VIP Works

It starts with a scale, say Likelihood, and five anchor statements that represent each position (e.g., "This is very likely to happen. I see strong factors supporting this outcome.").

You give it text:

"Very likely - I've been burned by bad survey tools, but this is different. Even if I validate later with humans, having directional answers this fast is worth it."

The system measures how closely the text matches the meaning of each anchor.

This captures nuance: a response might be 42% "Very Likely" and 37% "Uncertain" simultaneously.

This works better than asking for ratings directly. AIs asked to rate 1-5 hedge with safe middle responses. Free text first, numeric mapping second unlocks natural distributions.

I Made It Up's VIP Features

Create personas. Run interviews with VIP ratings. Get distributions and mean scores.

I Made It Up works entirely inside ChatGPT and Claude. No separate app.

Your AI working alongside you to analyse and derive qualitative insights (stories, themes, quotes) and quantitative patterns (distributions, comparisons, trends).

Current limits: 6 personas per group, 6 questions per interview. Need more? Custom work: sales@imadeitup.ai

Coming: Larger groups. Longer interviews. Statistical analysis built-in. Export to research tools.

How You Can Use VIP

Marketers

Three campaign concepts. Which one won't waste the budget?

Test all three in an hour. One's lukewarm. Another's strong positive. You launch the winner with confidence. Budget saved.

UX Researchers

Two weeks to recruit a handful of users. Or broad coverage in minutes.

Test first. Discover power users love it, casual users hate it. Opposite reactions. Redesign before recruiting real people. Validate a better design first time, not discover fundamental problems.

Brand Strategists

Client asks "how will the market respond?" You have two days.

Explore stakeholder perspectives today. Identify competitive perception gaps. Generate hypotheses about what drives differences. Present tomorrow with specific recommendations for validation work.

Human-Alignment

PyMC Labs and Colgate-Palmolive developed Semantic Similarity Rating, validating it against 9,300 real human participants across 57 consumer surveys (150-400 participants each).

The result? 90% correlation between synthetic and human response patterns. That's the theoretical maximum -- hitting the ceiling set by test-retest reliability.

It can replicate human demographic trends: age effects, income sensitivity, product category preferences. Aligned with real survey data at the distribution level.

VIP today implements the core mechanics but at smaller scale: 6 personas per interview and the ability to compare interviews, versus the 150+ used by the research. We're building toward that scale with real demographic data support on the roadmap.

Limitations

Sample size matters. Correlation with human responses requires 150+ participants. We currently support 6 personas per interview. Treat results as exploratory, not statistically rigorous.

Demographic validity. The validated research used real-world sourced demographic data. We use AI-generated demographics for flexibility, sacrificing validation rigor. Real demographic data available on request. Self-upload coming soon.

Knowledge gaps. Age and income effects replicated. Gender and regional patterns weaker. Validation used consumer products—niche domains may differ. Accuracy depends on AI knowledge of your demographic and subject.

Bias. AI learned from internet text, books and other media with embedded biases. Review outputs critically.

Not a replacement for human research. Use to explore and test hypotheses. Validate critical decisions with real humans.

Get Started

Get ChatGPT or Claude.

Go to imadeitup.ai and find the Getting Started instructions.

Bespoke enterprise-scale research? Talk to us.