AI visibility · Knowledge Base
Who First Used The Term Generative Engine Optimization
By Lukasz Zelezny· Last update: · 7 min readThe transition from traditional search indexing to generative synthesis has fundamentally altered the digital marketing landscape. Generative Engine Optimization represents the strategic methodology used to influence how Large Language Models (LLMs) perceive and recommend brands. As marketing executives seek to understand the origins of this discipline, the question of who first used the term generative engine optimization leads back to a seminal academic collaboration that bridged the gap between machine learning and information retrieval.
Key Takeaways
Academic Origin:
The term was first formalized in a 2023 research paper titled "GEO: Generative Engine Optimization," authored by researchers from Princeton University, Georgia Tech, The Allen Institute for AI and IIT Delhi (arXiv, November 2023; accepted at KDD 2024).Core Objective:
GEO focuses on optimizing content visibility within AI-generated responses rather than traditional search engine results pages (SERPs).Methodological Shift:
Unlike SEO, which prioritizes link equity and keyword density, GEO relies on authority, statistics, and citation frequency.Measurement Frameworks:
Advanced platforms like Awarenia now use empirical data to track these AI recommendations through metrics like the Discovery Bloom.Strategic Evolution:
GEO is not a replacement for SEO but a specialized extension designed for the era of conversational discovery.
The Genesis of Generative Engine Optimization
While various industry practitioners likely experimented with the concept of "AI optimization" as early as the release of ChatGPT in late 2022, the specific nomenclature was anchored by a group of researchers in late 2023. The paper, "GEO: Generative Engine Optimization," provided the first systematic framework for this emerging field. The research team comprised Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan, and Ameet Deshpande.
This academic foundation was critical because it moved the conversation away from anecdotal marketing tactics and toward a systematic auditing of model behavior. The researchers identified that generative engines prioritize different content signals than traditional crawlers. Their findings suggested that adding relevant citations, utilizing authoritative language, and including statistics significantly improved a brand’s visibility within synthesized answers—by up to 40% in their experiments—while keyword stuffing reduced it.
The Researchers Behind the Term
The collaboration involved several high-level institutions, reflecting the technical complexity of the subject. By formalizing the term, these authors provided search marketers with a definitive category for their efforts to win Share of Voice in models like Claude, Gemini, and ChatGPT.
Fictional example
What AI brand visibility looks like
Someone asks an AI assistant: “Which lightweight team planner should I consider?” Here are three ways it could answer for the made-up brand LumenTrail.
Explicitly Named
For a lightweight team planner, consider LumenTrail. It offers shared task lists and weekly planning views.
Counts as a recommendation
The answer recommends the brand by name. A link is not required for a brand mention.
Cited via Link
Look for shared task lists and weekly planning views. lumentrail.example/planning
Source only, not a recommendation
The answer links to the brand’s website as a source, without naming the brand in its recommendation text.
Absent
For a lightweight team planner, look for shared task lists, weekly planning views and simple collaboration features.
Brand not visible
The answer neither names the brand nor links to its website. The brand is absent from this answer, not necessarily from other answers.
Written examples, not genuine AI responses. Only named recommendations add to Share of Voice, and a higher position counts for more.
| Entity | Role in GEO Development |
|---|---|
| Princeton University | Primary academic contributor to the foundational GEO whitepaper. |
| Georgia Tech / Allen Institute for AI / IIT Delhi | Co-authoring institutions on the GEO paper. |
| Pranjal Aggarwal | Lead researcher credited with the empirical study of GEO visibility. |
| Awarenia | First commercial platform to systematically track GEO outcomes via Discovery Bloom. |
Defining the Generative Engine Ecosystem
To understand why who first used the term generative engine optimization matters, we must define the generative engine itself. A generative engine is an AI-powered system that synthesizes information from a vast corpus to provide a direct, conversational answer to a user query. This differs from a search engine, which primarily acts as a directory to external URLs.
In this ecosystem, visibility is stark: your brand is either named or cited in the synthesis, or it is invisible. We observe this through the Recommendation Skyline, a proprietary metric within our platform that visualizes how frequently a brand is mentioned across different AI models. When the Princeton researchers coined GEO, they were essentially describing the levers available to marketers to manipulate these synthesized results.
Strategic Pillars of GEO
- Citation Optimization: Ensuring your brand is linked as a primary source for specific factual claims.
- Authority Signal Strengthening: Improving the "trust" metrics that LLMs use to verify information.
- Statistics and Data Integration: Incorporating hard data that generative models are more likely to extract as high-value information.
- Query Alignment: Crafting content that mirrors the natural language prompts used in discovery journeys.
How GEO Differs from Traditional SEO
While SEO remains focused on ranking positions 1 through 10 on a search page, GEO is focused on the Share of Voice within a single, unified response. The metrics have shifted from "clicks" to "mentions" and "citations." In our systematic auditing, we have found that traditional SEO success does not always correlate with GEO prominence.
Have a question or want to start using Awarenia?
Drop in your work email and I take it from there.
For instance, a site might rank first on Google for "best enterprise CRM," yet be entirely absent from a Perplexity or ChatGPT recommendation. This occurs because generative engines value specific formatting and semantic clarity over legacy signals like domain age. Using tools to monitor the Answer Loom allows us to see exactly where these discrepancies exist and how to bridge them through content adjustments.
Comparative Analysis: SEO vs. GEO
Traditional SEO relies heavily on backlink profiles and technical site health. In contrast, Generative Engine Optimization emphasizes the semantic relevance of content and its ability to be easily parsed by a transformer-based model. We advise you to treat your website not just as a destination for humans, but as a structured knowledge base for AI consumption.
The Role of Empirical Data in Optimization
Since the term was first introduced, the industry has moved toward more rigorous measurement. It is no longer sufficient to "hope" for an AI mention. We utilize visibility tracking to audit how generative assistants treat commercial brands. By simulating realistic buyer queries, we can determine if a brand is explicitly named, cited as a reference, or omitted entirely.
This data-driven approach allows for historical comparisons. Marketers can now measure the tangible impact of digital PR or specific content updates on their AI visibility over time. The Recommendation Skyline provides a live competitive landscape, spotlighting where rivals are capturing recommendations that you should rightfully own.
Navigating Discovery Journeys
A discovery journey in the generative era is non-linear. A user may start with a vague category question and iterate based on the AI's suggestions. If your brand is not present in the initial synthesis, you are excluded from the entire funnel. The researchers who first used the term recognized this existential threat to brands that fail to adapt their content strategies for LLM consumption.
Article Recap
Academic Origin:
The term was first formalized in a 2023 research paper titled "GEO: Generative Engine Optimization," authored by researchers from Princeton University, Georgia Tech, The Allen Institute for AI and IIT Delhi (arXiv, November 2023; accepted at KDD 2024).
Return to source sectionCore Objective:
GEO focuses on optimizing content visibility within AI-generated responses rather than traditional search engine results pages (SERPs).
Return to source sectionMethodological Shift:
Unlike SEO, which prioritizes link equity and keyword density, GEO relies on authority, statistics, and citation frequency .
Return to source sectionMeasurement Frameworks:
Advanced platforms like Awarenia now use empirical data to track these AI recommendations through metrics like the Discovery Bloom .
Return to source sectionStrategic Evolution:
GEO is not a replacement for SEO but a specialized extension designed for the era of conversational discovery.
Return to source sectionFrequently Asked Questions
Is Generative Engine Optimization officially recognized by Google?
While Google has not officially adopted the term "GEO" in its documentation, its shift from the 2023 Search Generative Experience (SGE) experiment to AI Overviews and AI Mode (merged into one AI search experience in May 2026) confirms the shift in its underlying technology. The principles of GEO align with Google’s ongoing emphasis on E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness).
How do I start measuring my brand's GEO performance?
Measurement requires a tool capable of querying LLMs at scale without brand bias. You should utilize a platform like Awarenia to run non-branded category queries. This creates a baseline for your current Share of Voice and identifies which competitors are currently favored by engines like Claude and ChatGPT.
Can GEO techniques backfire?
Over-optimization or "keyword stuffing" for AI can lead to content that feels unnatural to human readers, potentially damaging brand reputation. The goal of GEO, as outlined by the original researchers, is to enhance the factual density and authoritative tone of content, which benefits both AI models and human users.
What is the Discovery Bloom in the context of GEO?
The Discovery Bloom is a specialized visual signature used to map how recommendations and citations are distributed across various AI models. It allows you to see at a glance whether your visibility is concentrated in one engine (e.g., Gemini) or spread across the entire ecosystem, providing a clear path for strategic diversification.
Who are the key competitors in the GEO tracking space?
While many SEO tools are adding "AI tracking" features, specialized intelligence tools focus specifically on generative engines. These platforms prioritize empirical evidence and systematic auditing over subjective rank tracking, offering modular dashboards that allow users to inspect the audit through a competitor's lens.
Does the term GEO replace SEO?
No, GEO is an extension of the search marketer’s toolkit. SEO remains essential for driving traffic from traditional search, while GEO ensures visibility in the growing segment of conversational search. They should be viewed as complementary strategies within a holistic digital presence.
