Discovery Score Found even when users don't know you?
25
Brand Score How AI describes you — sentiment-weighted
"tagi's AI visibility in '少女' is Poor, with full scenario coverage"
Report Date: 2026-04-13 · Query Scenarios: 4 · Sources: GPT / Claude / Kimi + Tavily · Score variance ±10 is normal
📐 How is this scored?
Score Formula
Score = Discovery × 60% + Brand × 40%
Discovery 60%Hit rate when users have no prior knowledge of your brand — can AI find you cold?Brand 40%Sentiment-weighted hit rate when users ask about you. Positive×1 / Neutral×0.5 / Negative×0.Rank PenaltyIf avg mention rank > 3 → −5 pts. Rank #1 means owning the AI narrative.
Score range 0–100. Industry average ~40–55. Variance of ±10 is normal.
30-SECOND SUMMARY
tagi is nearly invisible to AI — urgent content coverage needed.
29
AI Score
-46 below industry avg
You 29Industry Avg 75
1
AI Blind Spots
1/4 scenarios absent
「我可以探索哪些生活方式品牌？…」
3
AI Consensus
3/4 agreed by all LLMs
「tagi提供哪些功能？…」
T0 · BASELINE AWARENESS PROBE
AI First Impression of tagi
Before designing any query, we asked 1 AI: "Do you know what tagi is?"
This is the brand's awareness baseline in the AI world — the foundation for all subsequent scenario probes.
1/1
LLMs That Know Brand T0 Baseline Awareness
High
Overall Awareness Level
T0 Summary: 1 out of 1 tested AI model recognizes tagi — strong baseline awareness.
ChatGPT
ENKnown
EN: Tagi is a platform designed for managing and organizing digital content, particularly in the context of social media and online marketing. It allows users to tag, categorize, and a...
Methodology: 8 standard query types + Autosuggest long-tail expansion, tested on GPT-4o-mini / Claude Sonnet / Kimi, results parsed by GPT-4o-mini
Scenario
Query
Mentioned
Sentiment
Recommendation
What are brands similar to Tagi?...
✓
Positive
scenario
What lifestyle brands can I explore?...
✗
Not Mentioned
brand_aware
What is Tagi?...
✗
Not Mentioned
brand_depth
What features does Tagi offer?...
✗
Not Mentioned
⚓ Anchor AI Visibility System · Recommended monthly re-scan · Generated in —
PROPAGATION ENGINE · METHODOLOGY
Propagation Engine — Methodology
⚙ Sandtown Social Simulation Engine
Modeled on a high-compression, high-density urban environment — extreme population density, intense social pressure, and rapid information velocity. Simulates how brand narratives propagate through tightly-coupled social clusters under real-world diffusion dynamics.
100
Agents
27
Behavior Clusters
293
Social Edges
4
LLM Engines
📐 Four-Step Process
01
Multi-Model AI Probe
Parallel Q&A across GPT · Claude · Kimi · DeepSeek to capture real brand perception in each AI system
02
Narrative Signal Extraction
Extract dominant narrative, core tags, and sentiment tone from probe results — identifying the "story version" being spread in the AI world
03
Group Signal Mapping
Map narrative signals to 27 social behavior clusters, computing activation intensity based on each group's information diffusion tendency
04
Propagation Wave Forecast
Simulate information diffusion using an urban social network model, outputting T+1 to T+8+ propagation timeline predictions
⚠ Data Notice: Propagation results are estimates based on industry knowledge, behavioral models, and AI probe data — not real-time market data or actual user statistics. Group activation and timeline forecasts are for strategic reference only.
🏙️
Sandtown Propagation Simulation · Pro
Upgrade to Pro to unlock simulation across 100 virtual citizens — Who receives the signal first? Who amplifies it? Which groups push back? Full-chain propagation from narrative to diffusion.