Methodology & Architecture

The Science of
Answer Engine Optimization

Search is transitioning from ranked links to direct generative synthesis. Soma quantifies entity citations, buyer discovery intent, and competitive share of voice across LLM answer engines.

Start SimulationAPI Explorer

LLM Visibility Index (LVI) Formulation

Deterministic multi-factor scoring model across answer engines

Version 2.4 · Institutional
# Composite Visibility Metric (Scale: 0 – 100)
LVI = ∑ [ w_e × ( C_e × α + (1 - (rank-1)/K)× β ) ] × σ(sentiment)
w_eEngine weighting (Perplexity: 0.35, ChatGPT: 0.35, Gemini: 0.20, Copilot: 0.10)
C_e ∈ {0, 1}Binary entity citation verified in answer payload
rankOrdinal position among recommended solutions (Top-K = 5)
σ(sentiment)Contextual sentiment multiplier (0.80 to 1.15)
Pillar 01

Citation Probability

The frequency at which an engine includes your brand domain in its synthesized footnote references.

Pillar 02

Recommendation Rank

Whether your brand is proposed as the primary #1 solution or listed secondary to competing alternatives.

Pillar 03

Sentiment & Context

Contextual safety ensuring your product is recommended for its core strengths rather than flagged as an antipattern.

Query Archetypes

The 4 AEO Query Taxonomies

Generative AI searches follow conversational intent rather than keyword stuffing. Soma clusters buyer queries into four institutional archetypes.

40% Weight

Problem-Solution

Unbranded pain-point discovery when buyers ask how to resolve an operational challenge.

"How do high-velocity teams eliminate sprint planning lag?"
30% Weight

Category Comparison

Head-to-head architectural evaluations comparing feature parity, latency, and pricing tiers.

"Linear vs Jira for modern startup engineering squads"
15% Weight

Industry Workflow

Stack integration questions evaluating compatibility with GitHub, Slack, and SOC2 requirements.

"Best issue tracker integrating directly into GitHub PR workflows"
15% Weight

Brand Defense

Direct validation queries testing whether the engine accurately describes your pricing, features, and roadmaps.

"Is Linear free for open source and small startup teams?"

AEO vs Traditional SEO: The Paradigm Shift

AEO · Generative
Answer Engines (ChatGPT, Perplexity, Gemini)
Synthesis over 10 Blue Links:Generative models read dozens of sources and produce one concise verdict. Winning requires authoritative citation, not just keyword density.
Zero-Click Answer Capture:Over 60% of technical queries are answered without visiting a website. Brand presence is won or lost directly inside the AI dialogue.
Sentiment & Entity Graph:Models evaluate whether your brand is an industry standard or a legacy tool based on documentation semantic density.
SEO · Algorithmic
Search Crawlers (Google & Bing SERP)
Algorithmic Rank Positions:Traditional metrics focus on position #1–10 on fixed search result pages based on PageRank and backlink equity.
Static SERP Snippets:Standard meta descriptions and structured schema markup (JSON-LD) directing traffic to landing pages.
Keyword Volume Targeting:Optimizing pages for historical search volumes rather than conversational intent nuances.

Ready to see the pipeline in action?

Run a live brand lookup, inspect topic clusters, and audit multi-engine citation results in real time.

Run SimulatorView REST Endpoints