VS-AEO: A Vertical-Specific Answer Engine Optimization Framework for Local Business Visibility in Large Language Model Responses

We study how users seek local service providers through large language model (LLM) conversations and how query patterns differ by industry vertical. Using the WildChat-1M dataset (838,000 conversations, April 2023–April 2024), we propose the VS-AEO (Vertical-Specific Answer Engine Optimization) framework and introduce a four-stage Vertical Query Mining and Classification System (VQMCS) to systematically extract, classify, and analyze queries across six service verticals: Legal, Medical, Automotive, Roofing, HVAC, and Home Services. We identified 85,028 queries (10.1%) related to local service discovery. Legal services dominated with 44.6% of queries, followed by Medical (27.5%), Automotive (13.5%), Roofing (12.4%), HVAC (1.4%), and Home Services (0.6%). We introduce the Local Business Authority Score (LBAS) with initial theory-derived vertical-specific weights. Inter-annotator agreement reached Cohen's κ = 0.81.