The wiki · page 14 of 25

The reference desk for winning online.

Evergreen guides — kept current, backed by our own scan data, written in plain talk. Every article opens with the answer, not a preamble.

09 — AI & Search

What Is an Answer Engine?

An answer engine is a search system that responds to a question with a direct, synthesized answer instead of a list of links. Tools like Google's AI Overviews, ChatGPT, Perplexity, and Gemini read across many sources and generate a single reply, often citing a few of them. For local businesses, this shifts the goal from ranking ten blue links to being the trusted source an answer engine quotes, summarizes, or recommends.

UPDATED 2026-07-09 · READ →

What Is an AI Agent?

An AI agent is a software system that uses a language model to pursue a goal by planning steps and taking actions through tools, rather than just answering a single question. It can look things up, call APIs, fill forms, and make decisions in a loop until the task is done. For local businesses, agents can handle bookings, answer detailed customer questions, qualify leads, and automate routine back-office work with far less hand-holding than a basic chatbot.

UPDATED 2026-07-09 · READ →

What Is Prompt Engineering?

Prompt engineering is the practice of writing clear, well-structured instructions to get accurate, useful, and consistent results from an AI language model. It covers phrasing, context, examples, constraints, and formatting that guide the model toward the desired output. For businesses using AI in customer chatbots, content, or automation, good prompt engineering is the difference between reliable, on-brand responses and vague or incorrect ones, without changing the underlying model itself.

UPDATED 2026-07-09 · READ →

What Is AI Hallucination?

An AI hallucination is when a language model produces information that sounds confident and plausible but is actually false, invented, or unsupported. Because these models predict likely text rather than look up facts, they can fabricate names, prices, citations, or details. For businesses using AI in chatbots or content, hallucinations are a real risk: an assistant might invent a policy or quote a wrong price. Grounding, guardrails, and human review are the main defenses.

UPDATED 2026-07-09 · READ →

What Is Structured Data for AI?

Structured data for AI is machine-readable markup, usually schema.org in JSON-LD, that labels the facts on a web page so search engines and AI systems can understand them precisely. Instead of guessing from prose, a machine reads that this is a business name, that is a price, this is a review rating. For local businesses, structured data helps you earn rich results, feed accurate details into AI answers, and reduce misunderstandings about who you are and what you offer.

UPDATED 2026-07-09 · READ →

What Is Content Grounding in AI?

Content grounding is the practice of tying an AI model's answers to specific, trusted source material instead of letting it rely on memory. The system retrieves relevant passages from your real content and instructs the model to answer only from them, often citing the source. For businesses, grounding is the main way to make an AI chatbot accurate, quoting your true hours, prices, and policies, and to sharply reduce the confident fabrications known as hallucinations.

UPDATED 2026-07-09 · READ →

What Is a Model Knowledge Cutoff?

A model knowledge cutoff is the date after which an AI language model has no built-in knowledge, because its training data stops there. Events, prices, or facts that changed after the cutoff are unknown to the model unless it is given fresh information. For businesses, this matters because an AI might quote outdated details or miss recent changes. Connecting a model to live data through retrieval or search is how systems overcome the cutoff.

UPDATED 2026-07-09 · READ →

Brand Mention vs Backlink for AI Search?

A backlink is a clickable link from another website to yours, long a core SEO signal, while a brand mention is any reference to your business by name, linked or not. For traditional ranking, backlinks pass authority. For AI search and answer engines, unlinked brand mentions matter more than they used to, because AI systems weigh how often and how credibly your business is talked about across the web, not just who links to you.

UPDATED 2026-07-09 · READ →

What Is ChatGPT?

ChatGPT is an AI chatbot made by OpenAI that answers questions and holds conversations in plain language. It is built on a large language model, software trained on vast amounts of text to predict and generate natural-sounding responses, so you can ask it to explain a topic, draft an email, or write code and get a reply in seconds. Launched in late 2022, it helped popularize everyday AI. For local businesses, ChatGPT matters because customers increasingly ask it for recommendations, so being described accurately online affects whether you are mentioned.

UPDATED 2026-07-10 · READ →

What Is Claude AI?

Claude is an AI assistant built by Anthropic that answers questions, writes, analyzes documents, and holds conversations in natural language. Like other chatbots it runs on a large language model, but Anthropic emphasizes safety and reliability, training Claude with a method it calls Constitutional AI. Claude is known for handling long documents, careful reasoning, and a thoughtful, measured tone. People use it for writing, research, coding, and summarizing. For businesses, Claude matters as one of the AI assistants customers now use to ask questions and get recommendations.

UPDATED 2026-07-10 · READ →

What Is Perplexity AI?

Perplexity AI is an AI-powered answer engine that responds to questions with a written summary and, crucially, cites the web sources it used, shown as clickable numbered links. Unlike a chatbot that answers from memory, Perplexity searches the live web in real time, then uses a language model to synthesize a sourced answer, blending a chatbot's speed with search's transparency. People use it for research and quick, referenced answers. For businesses, being among Perplexity's cited sources is a new form of visibility that can send referral traffic and shape recommendations.

UPDATED 2026-07-10 · READ →

What Is Google Gemini?

Google Gemini is Google's family of AI models and the chatbot assistant built on them, designed to answer questions, generate text and images, write code, and reason across text, images, audio, and video. Gemini powers Google's standalone AI assistant and is woven into products like Search, where it generates the AI Overviews that appear above traditional results, plus Workspace apps such as Gmail and Docs. Because Gemini helps produce the AI summaries millions see in Google Search, how it represents a business directly affects local discovery and visibility.

UPDATED 2026-07-10 · READ →

What Is GPTBot?

GPTBot is OpenAI's web crawler, the automated bot that browses public web pages to gather text used for training and improving its AI models, including ChatGPT. It identifies itself with the user-agent "GPTBot" and obeys robots.txt rules, so site owners can allow or block it. Allowing GPTBot can help your content surface in AI answers; blocking it keeps your pages out of OpenAI's training data. The choice is yours, per-path, and reversible at any time.

UPDATED 2026-07-10 · READ →

What Is ClaudeBot?

ClaudeBot is Anthropic's web crawler, the automated bot that reads public web pages to help build and improve its Claude AI models. It identifies itself with the user-agent "ClaudeBot" and respects robots.txt, so site owners can allow or block it per path. Allowing ClaudeBot lets your content inform Claude's knowledge and potentially be referenced in answers; blocking keeps your pages out of Anthropic's crawl. Like other AI crawlers, the choice is yours and can be changed at any time.

UPDATED 2026-07-10 · READ →

What Is PerplexityBot?

PerplexityBot is the web crawler used by Perplexity, an AI answer engine that responds to questions with summaries and cited sources. It reads public pages so Perplexity can index them and link to them when they support an answer. PerplexityBot identifies itself in the user-agent and respects robots.txt, so you can allow or block it. Because Perplexity cites the sources behind its answers, allowing PerplexityBot gives your pages a chance to appear as a linked citation and earn referral traffic.

UPDATED 2026-07-10 · READ →

What Is an AI Crawler?

An AI crawler is an automated bot that browses public web pages to collect text for artificial-intelligence systems, either to train large language models or to retrieve up-to-date content for AI answers. Examples include GPTBot, ClaudeBot, and PerplexityBot. AI crawlers work like search-engine crawlers mechanically, requesting pages and following links, and most respect robots.txt so site owners can allow or block them. They differ from search crawlers mainly in purpose: powering AI assistants rather than building a traditional search index.

UPDATED 2026-07-10 · READ →

What Is a Token in AI?

A token is the basic unit of text an AI language model reads and generates. Models do not process whole words or characters directly; they break text into tokens, which are often word fragments, whole short words, or punctuation. A rough rule of thumb is that one token equals about four characters of English, or roughly three-quarters of a word. Tokens matter because they determine an AI's cost, how much text fits in its context window, and how long responses can be.

UPDATED 2026-07-10 · READ →

What Is a Context Window?

A context window is the maximum amount of text, measured in tokens, that an AI language model can consider at one time. It covers both the input you provide and the output the model generates, so a long prompt leaves less room for the response. Anything beyond the window is invisible to the model, which is why very long documents or conversations can lose earlier detail. A larger context window lets a model reason over more material at once, improving accuracy on big or complex tasks.

UPDATED 2026-07-10 · READ →

What Is Fine-Tuning in AI?

Fine-tuning is the process of taking a pre-trained AI model and training it further on a smaller, focused dataset so it performs better on a specific task, tone, or domain. Instead of building a model from scratch, you adjust an existing one with your own examples, teaching it patterns it should follow. Businesses use fine-tuning to make a model consistently match a brand voice, follow a format, or handle a niche subject. It is powerful but not always necessary; often good prompting or retrieval solves the problem more cheaply.

UPDATED 2026-07-10 · READ →

What Is a System Prompt?

A system prompt is a set of hidden instructions given to an AI assistant before a conversation begins, defining its role, tone, rules, and boundaries. Unlike the messages a user types, it is set by whoever builds the AI feature and shapes how the model behaves throughout the chat. It might tell the assistant to act as a friendly support agent, stay on certain topics, or avoid giving certain advice. A well-written system prompt is one of the most powerful, low-cost ways to control an AI's behavior.

UPDATED 2026-07-10 · READ →

What Is Multimodal AI?

Multimodal AI is artificial intelligence that can understand and work with more than one type of input or output, such as text, images, audio, and video, together rather than in isolation. A multimodal model can look at a photo and describe it, read a chart and answer questions about it, transcribe speech, or combine a text instruction with an image. This contrasts with older single-mode models that handled only text. For businesses, multimodal AI enables tools like visual product search, image-based support, and richer, more natural assistants.

UPDATED 2026-07-10 · READ →

What Is Conversational Search?

Conversational search is a way of searching using natural, spoken-style language and back-and-forth follow-up questions, rather than typing short keywords. Instead of "plumber near me open Sunday," a person might ask "who can fix a burst pipe in my area today?" and then refine with "one with good reviews under a hundred dollars." Powered by AI assistants and voice tools, conversational search understands intent and context across a dialogue. It changes SEO strategy by rewarding content that answers real questions clearly rather than just matching keywords.

UPDATED 2026-07-10 · READ →

What Is Voice Search?

Voice search is the act of speaking a query to a device instead of typing it, then getting a spoken or on-screen answer. It runs through assistants like Google Assistant, Siri, Alexa, and phone keyboards, which convert speech to text, interpret intent, and return a result. Queries tend to be longer, conversational, and often local, such as "who's the best plumber near me open now." For small businesses, ranking for these spoken, question-shaped searches means structured content, fast pages, and an accurate Google Business Profile.

UPDATED 2026-07-10 · READ →

What Is a Chatbot Widget?

A chatbot widget is the small chat box that appears in the corner of a website, letting visitors ask questions and get instant answers. It loads as a bubble that expands into a conversation panel, powered by scripted rules, an AI model, or a live agent handoff. Businesses use it to answer common questions, qualify leads, book appointments, and capture contact details around the clock. Modern widgets often connect to a knowledge base or your own content so answers stay accurate, and route complex issues to a human when needed.

UPDATED 2026-07-10 · READ →

What Is AI Content Detection?

AI content detection is the use of tools that try to estimate whether text was written by a human or generated by an AI model. Detectors analyze statistical patterns like predictability and sentence uniformity, then output a probability score rather than a definite verdict. They are used by schools, publishers, and hiring teams, but accuracy is limited and false positives are common. For businesses, the bigger lesson is that search engines reward helpful, accurate content regardless of how it was produced, so quality and originality matter more than passing a detector.

UPDATED 2026-07-10 · READ →

What Is an AI Overview Source?

An AI Overview source is a website that Google cites in the AI-generated summary shown at the top of many search results. When Google's AI answers a query, it pulls facts from a handful of trusted pages and links them as sources beside or beneath the summary. Being chosen sends referral traffic and signals authority. Selection favors pages that clearly and directly answer the question, demonstrate real expertise, load fast, and use clean structure and schema. It is not a paid placement; it is earned through helpful, well-organized, trustworthy content.

UPDATED 2026-07-10 · READ →

What Is Bing Copilot?

Bing Copilot is Microsoft's AI answer assistant built into Bing search and Windows, which responds to questions with conversational, cited summaries instead of just a list of links. It combines a large language model with live web results, so answers reflect current information and link to sources. Users can ask follow-ups, generate text and images, and get help with tasks. For businesses, Copilot is another AI surface that cites websites, so appearing as a source depends on the same fundamentals as Google: clear, authoritative, well-structured content that is easy for AI systems to read and trust.

UPDATED 2026-07-10 · READ →

What Is Grounding Data in AI?

Grounding data is the trusted, external source information an AI model consults to keep its answers factual and current, instead of relying only on what it memorized during training. By retrieving relevant documents, databases, or web pages and answering from them, a grounded system reduces hallucinations and can cite where facts came from. It is the backbone of retrieval-augmented generation, on-site AI chatbots, and AI search answers. For businesses, grounding an AI on your own verified content, services, policies, and FAQs, means it gives accurate, on-brand answers instead of confidently inventing details.

UPDATED 2026-07-10 · READ →

What Is an Embedding Model?

An embedding model is an AI model that turns text, or other data, into a list of numbers called a vector, where similar meanings produce similar vectors. This lets computers compare content by meaning rather than exact words, which powers semantic search, recommendations, and retrieval-augmented generation. When you ask a grounded chatbot a question, an embedding model converts both your query and the source content into vectors so the system can find the closest matches. For businesses, embeddings are the quiet engine behind AI features that understand intent instead of just matching keywords.

UPDATED 2026-07-10 · READ →

What Is a RAG Pipeline?

A RAG pipeline is the retrieve-then-generate workflow that lets an AI answer questions using your own content instead of only its training. RAG stands for retrieval-augmented generation. When a user asks something, the pipeline retrieves the most relevant passages from your indexed documents, then feeds them to a language model that composes an answer grounded in that material, often with citations. This keeps responses accurate, current, and specific to your business, and sharply reduces hallucinations. It is the standard architecture behind grounded chatbots, internal knowledge assistants, and AI search that answers from a controlled set of trusted sources.

UPDATED 2026-07-10 · READ →

Looking for news, case studies, and opinion pieces? That's the blog →