AI-Search Growth System: A Methodology for Generative Engine Optimisation at Scale

Large language models (LLMs) such as ChatGPT, Gemini, Claude and Perplexity now surface commercial recommendations directly inside their responses, creating a new ranking surface that conventional search-engine optimisation (SEO) tools do not measure. This paper describes the methodology behind Zaiq's AI-Search Growth System: a production pipeline that (1) programmatically queries four frontier LLMs to measure brand-mention frequency and rank position across a target keyword set, (2) identifies the structural and semantic gaps that suppress a brand from appearing in LLM-generated answers, and (3) engineers targeted content and citation assets to close those gaps. We document the prompt-construction strategy, the normalisation scheme used to make cross-model rank scores comparable via the Cross-Model Rank Score (CMRS) metric, the iterative optimisation loop, and the evaluation framework. The system has been deployed for live South African businesses including a campaign across 28 South African restaurants, demonstrating measurable rank uplift within a single working day from initial audit to shipped fix. Zaiq is an AI engineering studio based in Johannesburg, South Africa (zaiq.co.za). This methodology is released openly under CC BY 4.0 to advance transparent practice in Generative Engine Optimisation (GEO).