How Our Recommendation Works
Learn how AI Finder selects the best fit from 22 AI platforms using a transparent matching process.
Process Overview
1. Profile Analysis
6 questions assess goals, skill level, budget, and environment
2. Match Scoring
Tag-based scores + bonuses + synergy combined
3. Personalized Picks
Top 5 recommendations with clear explanations
Tag-Based Matching System
At the heart of AI Finder is a sophisticated tag matching system. Each AI platform has unique tags representing its characteristics, and your answers are also converted into tags.
How Tag Matching Works
Each option you select in the 6 questions is linked to specific tags. For example, selecting 'Write code' adds the 'code-generation' tag, while 'Ease of use' adds the 'ease-of-use' tag. These collected tags are then compared against each AI platform's tags to calculate match scores.
Tag Categories
Core Features
HighPrimary AI capabilities like text generation, code generation, image creation, voice processing
Platform Traits
MediumDeveloper-friendly, enterprise-focused, multimodal support, etc.
Usage Environment
MediumWeb, mobile, API, office integration, etc.
Budget
LowFree, low-cost, mid-range, premium pricing tiers
Scoring System
We don't just count tag matches. Our multi-layered scoring system ensures accurate, nuanced recommendations.
Score Components
Base Match Score
When user tags match platform tags, points are awarded based on tag weight. Core feature tags (e.g., code-generation) carry 15-20 points, while environment tags (e.g., mobile) carry 6-8 points.
Specialization Bonus
Each platform has unique strengths. GitHub Copilot gets bonuses for coding tags, Midjourney for image generation tags. This ensures platforms with clear domain expertise score higher in their specialty.
Synergy Bonus
When related tags match together, bonus points are added. For example, if 3+ development tags (development, code-generation, api-access) all match, a synergy bonus applies—rewarding platforms that offer comprehensive development environments.
Constraint Penalties
If a platform fails to meet your requirements, penalties apply. For example, if you need 'local installation only' but the platform is cloud-only, its score is reduced.
Final Score Calculation
Final Score = (Base Match Score + Specialization Bonus + Synergy Bonus - Penalties) × Normalization Factor. This score is normalized to 0-100 and displayed as a match percentage.
Personalization Logic
Even with identical answer combinations, recommendations can vary based on your overall profile.
User Type Classification
Based on your test results, you're classified into one of 7 user types (Maker, Idea, Organizer, Explorer, Safety-first, Natural Communicator, Casual). This classification is reflected in the recommendation explanation, providing personalized context for why each platform fits you.
Context-Aware Recommendations
We consider not just features, but your entire context—usage environment (personal/work/creative), skill level, and budget. For example, with the same 'image generation' goal, a professional designer might be recommended Midjourney, while a general user might get Canva.
Reliability & Transparency
AI Finder follows these principles to provide accurate, trustworthy recommendations.
Data-Driven Evaluation
Each platform's tags and characteristics are defined using official documentation, real usage experience, and user feedback.
Regular Updates
AI platforms evolve rapidly. We periodically update our data to reflect new features, price changes, and service updates.
Unbiased Recommendations
We're not influenced by partnerships or ad revenue. Recommendations are based purely on your requirements and platform characteristics.
Transparent Results
We show why each platform was recommended—matched tags and scores are visible. You can always verify the reasoning behind your results.
Data Management
We systematically manage platform data to ensure accurate recommendations.
Platforms
We continuously add major AI platforms.
Update Frequency
New features, price changes, and updates are reflected regularly.
Tags
Granular tags enable precise matching.
If your recommendation doesn't match your actual experience, let us know through Discussion or Contact. Your feedback helps us improve the algorithm.