AI search comes with its own dialect, and most of it is jargon dressed up to sound complicated. You do not need a marketing degree to follow it. You need a short list of AI search terms, defined honestly, so you can read a proposal, evaluate an agency, or just understand why a competitor keeps getting recommended by ChatGPT and you do not.
That is what this glossary is for. We have organized it the way the concepts actually connect rather than alphabetically, so by the end you will understand not just what each term means but how the pieces fit together to decide who an answer engine names. If you want the deeper version of any concept, we have linked out where it helps.
The core AI search terms
Start here. These five terms are the backbone of every conversation about getting found by AI, and almost everything else in this glossary builds on them.
Answer engine
An answer engine is any tool that responds to a question with a direct answer instead of a list of links. ChatGPT, Perplexity, Google Gemini, Google AI Overviews, and Microsoft Copilot are all answer engines. The shift matters because a person asking "who is the best mortgage broker near me" used to get ten options to compare. Now they often get one recommendation, and the businesses that are not named simply do not exist in that moment.
AEO (answer engine optimization)
AEO is the practice of structuring your website, content, and reputation so answer engines pick your business as the answer. It is the discipline our whole company is built around. If you only learn one term from this AI search glossary, make it this one, because it is the goal everything else serves. Our full guide to answer engine optimization walks through the strategy end to end.
GEO (generative engine optimization)
GEO is a closely related term focused on getting your content quoted and cited inside AI-generated answers. Where AEO emphasizes being the recommended choice, GEO emphasizes being the source. In day-to-day work the line blurs, and you will see the two used almost interchangeably. We treat them as two angles on the same objective: being the business the model trusts, names, and links to.
LLM (large language model)
A large language model is the AI system underneath an answer engine. GPT (which powers ChatGPT), Gemini, and Claude are all LLMs. They are trained on enormous amounts of text and learn patterns about how the world is described. That detail matters for your business: an LLM does not "know" you are excellent unless that excellence is described, repeatedly and consistently, in the text it learned from.
LLMO (large language model optimization)
LLMO is yet another label for the same family of work, emphasizing the language model itself rather than the search experience around it. AEO, GEO, and LLMO are not three different services you need to buy separately. They are overlapping names for the work of becoming the answer AI gives.
How answer engines actually decide
Knowing the vocabulary is only useful if you understand the mechanics behind it. Across the audits we run, the businesses that win in AI search are rarely the flashiest. They are the ones the models can describe confidently because the evidence is everywhere and it all agrees.
Training data
Training data is the body of text an LLM learned from before it was released. If your business was barely mentioned online when a model was trained, the model has little to say about you. This is why a quiet, well-run company can be invisible to AI while a louder competitor gets named.
Retrieval and grounding
Modern answer engines do not rely on training data alone. They also retrieve live information from the web at the moment of the question and use it to "ground" the answer in current sources. This is good news: it means you can earn visibility now without waiting for the next training run, as long as the current web tells a clear, consistent story about you.
Citation
A citation is when an answer engine names or links to your site as a source. Citations build authority over time and can send referral clicks your way. They are earned by publishing clear, answer-first content and by being mentioned on the directories and review platforms the models already lean on.
Hallucination
A hallucination is when an AI states something false with confidence, including wrong facts about your business such as outdated hours, the wrong service area, or a phone number that is not yours. The defense is the same as the offense: give the models accurate, structured, consistent information so they have no reason to guess. Our explainer on how AI assistants decide who to recommend goes deeper on the decision process.
The technical building blocks
A handful of technical terms come up constantly because they are the levers you can actually pull. None of them require you to be an engineer to understand.
Schema markup (structured data)
Schema markup is code added to your pages that labels what the content means: your business name, services, location, reviews, prices, and FAQs. It turns a paragraph a human reads into facts a machine can extract cleanly. It does not buy a recommendation, but it removes ambiguity, and ambiguity is what gets businesses skipped or misquoted.
llms.txt
An llms.txt file is a plain-text file at the root of your site that hands AI crawlers a tidy summary and map of your most important pages. Think of it as a welcome guide written for machines. It will not rescue a weak presence, but it lowers friction for the models that read it. For a fuller treatment, see our piece on what llms.txt is and whether your business needs one.
NAP consistency
NAP stands for name, address, and phone number. NAP consistency means those details match exactly everywhere they appear: your site, your Google Business Profile, directories, and review sites. Inconsistent NAP data is one of the most common, and most fixable, reasons a business confuses the models and loses the recommendation.
Google Business Profile
Your Google Business Profile is the free listing that controls how you appear in Google Maps and local results, and it feeds the data many answer engines lean on for local recommendations. A complete, accurate, actively reviewed profile is one of the highest-leverage assets a local business has in AI search.
Terms that describe the new behavior
Finally, a few terms describe how customer behavior and traffic are changing, which is the context that makes the rest of this glossary matter.
Zero-click search
Zero-click search is when a person gets their answer directly from the results and never clicks through to a website. Answer engines accelerate this. The practical takeaway is that being named in the answer is now its own form of marketing, even when the click never happens.
AI Overviews and AI Mode
AI Overviews are Google's AI-generated summaries that appear above the traditional results. AI Mode is Google's fuller conversational search experience. Both decide which businesses get summarized and surfaced, which is why they belong in any business owner's vocabulary.
AI visibility
AI visibility is the umbrella measure of how often, and how favorably, answer engines mention your business. It is the scoreboard for everything above. The first step in any engagement we run is simply measuring it, because you cannot improve what you have never checked.
A quick-reference table
For when you just need the one-line version:
| Term | What it means in one line |
|---|---|
| AEO | Optimizing to be the business AI recommends |
| GEO | Optimizing to be the source AI quotes and cites |
| LLM | The AI model behind ChatGPT, Gemini, and Claude |
| Answer engine | A tool that answers questions instead of listing links |
| Citation | An AI naming or linking to you as a source |
| Schema | Code that labels what your content means for machines |
| llms.txt | A plain-text map of your site for AI crawlers |
| Zero-click | Getting the answer without visiting a website |
How the pieces fit together
Here is the chain that connects every term above. Answer engines run on LLMs. LLMs describe your business based on what the web says about you, supported by structured data they can read cleanly. When that description is clear, consistent, and corroborated by reviews and citations, the model recommends you. AEO, GEO, and LLMO are just names for the work of making that chain hold.
You do not have to master all of it at once. The businesses we see win usually start with the basics that compound:
- Lock down NAP consistency and a complete Google Business Profile.
- Add schema markup so your facts are machine-readable.
- Publish answer-first content that responds directly to real questions.
- Earn reviews and citations on the platforms the models already trust.
- Measure your AI visibility so you know whether any of it is working.
None of this is theoretical. We watched a Seattle mortgage broker go from invisible in AI search to the number-one AI-recommended broker in his market, with roughly 30 leads and four closed deals inside six weeks, by getting these fundamentals right rather than chasing tricks.
Keep this glossary handy. The terminology will keep shifting as the tools evolve, but the underlying idea is stable: AI recommends the business it can understand and trust. Get clear, get consistent, and get measured, and most of the jargon stops being intimidating and starts being useful.