The eesel AI glossary

The percentage of customers who leave a support queue, chat, or call before reaching an agent or getting an answer.

Software that supports a human agent in real time by surfacing answers, drafting replies, and pulling context while the person stays in control of the response.

AI that acts with autonomy, planning its own steps, using tools, and adapting to results to reach a goal, rather than producing a single response to a single prompt.

An AI system that pursues a goal on its own, understanding a request, deciding what steps to take, calling tools, and completing the task rather than just answering one question.

A software tool that uses generative AI to research a topic and draft a complete, structured blog post from a brief or keyword.

A software application that uses artificial intelligence to hold a natural-language conversation, understanding what a user means and generating a relevant reply.

An AI assistant that gives customers personalized, proactive guidance across their whole journey, anticipating needs and handling requests end to end rather than only answering when asked.

AI content detection is the use of software that estimates the probability a piece of text was generated by an AI model rather than written by a human.

The use of generative AI models to produce written, visual, or audio content from a prompt, brief, or set of source material.

An AI assistant that works alongside a person inside their existing workflow, suggesting, drafting, and surfacing information while the human stays in control of the decision.

The use of artificial intelligence to understand, respond to, and resolve customer questions, either on its own or by assisting human agents.

Tying an AI's answers to a trusted source of truth, like your help center, docs, and past tickets, so it responds with your facts instead of inventing them.

The rules, limits, and safety checks placed around an AI system to keep its behavior within intended, safe, and accurate bounds.

When an AI model generates an answer that sounds confident and fluent but is factually wrong or entirely made up.

The coordination of multiple AI models, tools, and steps into a single workflow so they work together toward one outcome.

Ticket summarization is the act of condensing a long support conversation into a short, accurate recap of what happened and what is still needed.

A tool that uses AI to help a person write and edit, suggesting text, rephrasing, fixing errors, and shaping drafts as they work.

Alt text is a written description of an image, added in HTML, that screen readers announce and search engines read when the image cannot be seen.

The visible, clickable words in a hyperlink, which tell readers and search engines what the linked page is about.

The practice of structuring content so AI answer engines and search features can extract a direct answer and cite it, rather than just rank a page in a list of links.

The automatic sorting, tagging, prioritizing, and routing of incoming requests so each one reaches the right place without manual review.

A support interaction that is fully closed out by software, with no human agent touching it, from the customer's first message to the resolved outcome.

The average total time an agent spends resolving a single contact, including talk time, hold time, and after-contact work.

A backlink is a hyperlink on one website that points to a page on another website.

A canned response is a pre-written reply that support agents can insert into a conversation to answer a common question quickly and consistently.

A canonical URL is the version of a page that a site declares as the master copy when the same or similar content is reachable at more than one address.

A software application that simulates a conversation with a user through text or voice, answering questions and guiding them through tasks.

Click-through rate is the percentage of people who click a link, ad, or search result out of the total number who saw it.

A confidence score is a number, usually between 0 and 1, that an AI model assigns to a prediction or answer to express how likely it is to be correct.

A contact center is a centralized team and system that manages customer interactions across multiple channels such as phone, email, chat, and social.

The percentage of conversations an automated channel handles to completion without escalating to a human agent.

A document that tells a writer what a piece of content should cover, including topic, target keyword, audience, structure, and angle.

A content calendar is a schedule that maps out what content a team will publish, when, and on which channels.

Content decay is the gradual decline in a page's traffic and search rankings over time after it once performed well.

A content gap is a topic or question your audience is searching for that your existing content does not adequately cover.

Content marketing is the practice of attracting and retaining an audience by creating useful content rather than directly advertising a product.

A content refresh is the process of updating an existing published page to restore or improve its accuracy, relevance, and search rankings.

The rate at which an organization plans, produces, and publishes new content over a given period.

A context window is the maximum amount of text, measured in tokens, that a language model can consider at once when generating a response.

Technology that lets software hold natural, back-and-forth conversations with people through text or voice.

Core Web Vitals are a set of standardized metrics from Google that measure a web page's loading speed, interactivity, and visual stability as experienced by real users.

Cost per ticket is the average total cost a support team spends to handle one customer request, found by dividing total support costs by the number of tickets resolved.

Crawl budget is the number of pages a search engine will crawl on a given site within a certain timeframe.

A metric that measures how satisfied customers are with a specific interaction, product, or service, usually from a short rating survey.

A short questionnaire that asks customers to rate how satisfied they were with a specific interaction, product, or service.

The rate at which customers stop doing business with a company over a given period, usually expressed as a percentage of the starting customer base.

A metric that measures how much effort a customer had to put in to get their issue resolved or their goal met.

Customer experience is the sum of every interaction a person has with a company across the entire relationship, and how those interactions make them feel.

The full path a customer takes with a company across every stage and touchpoint, from first awareness through purchase, onboarding, support, and renewal.

The total revenue a business can expect from a single customer across the entire length of their relationship.

The ability of a company to keep its existing customers over time, usually measured as the percentage of customers who remain across a given period.

Customer satisfaction is a measure of how well a product, service, or interaction meets a customer's expectations.

Support resources that let customers find answers and resolve their own issues without contacting a human agent.

Customer service is the help a company provides to people before, during, and after they buy, so they can use the product and resolve any problems.

The use of technology to handle customer service tasks with little or no human effort, from routing and tagging tickets to answering and resolving them.

Customer support is the help a company gives customers to solve problems and answer questions about a product or service they already use.

A subset of machine learning that uses neural networks with many layers to learn complex patterns directly from raw data like text, images, and audio.

The percentage of incoming customer questions resolved through self-service or automation without being handled by a human agent.

A third-party score that predicts how well a whole website is likely to rank in search results, based mostly on its backlink profile.

The length of time a visitor stays on a page after clicking through from search before returning to the results, used as a signal of how well the page met their intent.

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness, the qualities Google's raters use to judge content quality.

Embeddings are numerical representations of text, images, or other data as vectors, where items with similar meaning sit close together in the vector space.

The act of moving a support request to a higher tier, a specialist, or a human when the current handler cannot or should not resolve it.

Escalation rate is the percentage of support interactions that get handed off to a higher tier, a specialist, or a manager instead of being resolved at first contact.

Evergreen content is material that stays relevant and useful for a long time after publishing, drawing steady traffic without needing frequent rewrites.

A short answer a search engine pulls from a web page and displays at the top of results, above the standard listings.

Few-shot learning is when a model learns to perform a task from a small number of examples shown directly in the prompt, rather than from a large training dataset.

Fine-tuning is the process of taking a pre-trained machine learning model and training it further on a smaller, task-specific dataset to adapt its behavior.

The share of support requests that are fully resolved in a single interaction, with no follow-ups, callbacks, or escalations.

First response time is how long a customer waits between submitting a request and receiving the first human or automated reply from a support team.

A foundation model is a large AI model trained on broad data that can be adapted to many different downstream tasks rather than built for just one.

A category of AI that creates new content such as text, images, audio, or code by learning patterns from large training datasets.

The practice of optimizing content so that generative AI engines surface, synthesize, and cite it when they assemble an answer from multiple sources.

The software a team uses to receive, organize, track, and resolve customer or employee support requests in one central place.

Hub-and-spoke content is a structure where one broad hub page links out to many focused spoke pages on related subtopics, all interlinked.

A design pattern where a person reviews, approves, or corrects an automated system's output at key points, keeping human judgment in the decision.

Indexing is the process by which a search engine analyzes a crawled page and stores it in its database so it can appear in search results.

The task of assigning a piece of text to one of a set of predefined intent labels that describe what the person is trying to do.

The process of figuring out the underlying goal behind a person's message, so a system can respond to what they actually want.

An automated phone system that interacts with callers through voice prompts and keypad input, routing them or answering before a human agent picks up.

Internal linking is the practice of connecting pages within the same website using hyperlinks.

Keyword difficulty is a score that estimates how hard it would be to rank on the first page of search results for a given keyword.

The process of finding and analyzing the search terms people type into search engines, to decide what content to create and how to rank for it.

A knowledge base is an organized, searchable collection of articles and documentation that lets people find answers to recurring questions on their own.

Knowledge management is the practice of capturing, organizing, sharing, and maintaining an organization's information so the right people can find and use it.

A type of AI model trained on huge amounts of text that predicts language well enough to understand and generate human-like writing.

Link building is the practice of getting other websites to link to your pages in order to improve their authority and search rankings.

A listicle is an article structured as a numbered or bulleted list, where each item is a self-contained entry under its own heading.

Live chat is a real-time messaging channel that lets customers and support agents converse instantly through a website or app.

Long-form content is in-depth written content, usually well over a thousand words, that covers a topic thoroughly rather than briefly.

A long-tail keyword is a longer, more specific search phrase that has lower search volume but usually clearer intent and less competition than a broad head term.

A branch of AI where systems learn patterns from data and improve at a task over time, instead of following rules written by hand for every case.

A macro is a saved set of actions a support agent can apply to a ticket in one click, combining a pre-written reply with updates like status, tags, or assignment.

A short HTML attribute that summarizes a web page and often appears as the snippet beneath the title in search results.

Structured information that describes a web page to search engines, browsers, and social platforms rather than to the reader directly.

An open standard that defines how AI applications connect to external tools and data sources through a consistent interface.

AI that can understand and generate more than one type of data, such as text, images, audio, and video, within a single model.

The branch of AI that turns structured data or an internal representation into fluent, human-readable text.

The field of AI focused on letting computers understand, interpret, and generate human language, from reading a support ticket to writing a reply.

The branch of AI focused on reading human language and extracting its meaning, intent, and structure so a machine can act on it.

A loyalty metric that measures how likely customers are to recommend a company, based on a single 0 to 10 question.

Off-page SEO is the set of signals from outside your own website that shape how search engines judge its authority and trustworthiness.

A support approach that connects every channel a customer might use into one continuous conversation with shared context.

The practice of optimizing the content and HTML elements of an individual web page, like its title, headings, copy, and structure, to rank higher in search.

A support outcome where a customer's issue is fully resolved in a single interaction, with no follow-ups, transfers, or repeat contacts.

Organic traffic is the visitors who arrive at a website through unpaid search engine results rather than through ads or other paid placements.

Outcome-based pricing is a model where the buyer pays for results actually delivered, such as resolved tickets, rather than for seats, licenses, or a flat subscription.

A pillar page is a comprehensive page covering a broad topic at a high level, linked to and supported by more detailed cluster pages.

Proactive support is the practice of reaching out to customers about an issue or need before they have to contact you about it.

Programmatic SEO is the practice of generating large numbers of search-targeted pages from a template and a structured data set.

The practice of designing and refining the instructions given to a language model so it produces accurate, useful, reliable output.

An attack where crafted input tricks a language model into ignoring its original instructions and following the attacker's instead.

The support-applied form of retrieval-augmented generation, where an AI fetches relevant content from a company's own knowledge before generating a customer-facing answer.

Reopen rate is the percentage of resolved support tickets that get reopened because the customer's issue was not actually solved.

The share of support requests that end in a confirmed solution, calculated as resolved tickets divided by total tickets over a period.

The total time it takes to fully resolve a customer support request, measured from when the ticket is opened to when it is closed.

A technique that retrieves relevant documents at query time and feeds them to a language model so its answer is grounded in those sources.

The practice of improving a website so it ranks higher in unpaid search results and attracts more relevant visitors.

The underlying goal a person has when they type a query into a search engine, usually grouped as informational, navigational, commercial, or transactional.

The average number of times people enter a given query into a search engine over a set period, usually expressed as monthly searches.

Semantic search is a search method that ranks results by meaning and intent rather than exact keyword matching, so it can return relevant items that share no words with the query.

Sentiment analysis is the use of language processing to detect the emotional tone of a piece of text, usually as positive, negative, or neutral.

The page a search engine returns after a query, listing organic results, ads, and rich features like snippets, images, and answer boxes.

A service desk is the single point of contact between an organization and its users for logging, tracking, and resolving requests and incidents.

A shared inbox is a single email mailbox that multiple team members can access and manage together to handle incoming messages collaboratively.

A routing method that matches each support request to the agent whose specific skills, language, or expertise best fit it, rather than the next available person.

A service level agreement is a documented commitment to deliver a service to a defined standard, such as replying to or resolving support requests within a set time.

Standardized code added to a web page that describes its content to search engines in a machine-readable format.

A suggested reply is an AI-drafted response shown to a support agent, who can edit, approve, or discard it before it reaches the customer.

The use of software to perform support tasks automatically, reducing the manual work involved in triaging, answering, and resolving customer requests.

The process of reviewing support interactions against a defined standard to measure and improve the quality of the service customers receive.

A support queue is the ordered list of customer requests waiting to be picked up and worked by a support team.

Technical SEO is the practice of optimizing a website's infrastructure so search engines can crawl, render, and index its pages efficiently.

A ticket backlog is the accumulated set of open support requests that have not yet been resolved or closed at a given point in time.

The practice of resolving a customer's question before it becomes a tracked support ticket, usually through self-service content or an automated answer.

The process of directing an incoming support request to the right person, team, or queue based on its content, priority, or other attributes.

Ticket tagging is the practice of attaching labels to support tickets so they can be sorted, routed, measured, and reported on.

Ticket volume is the total number of support requests a team receives over a given period, used to size workload, staffing, and demand.

A ticketing system is software that turns each incoming request into a trackable ticket and manages it through to resolution.

An HTML element that sets a web page's title, used as the clickable headline in search results and the label on a browser tab.

A token is the basic unit of text a large language model reads and generates, usually a word, part of a word, or a single character.

Tone of voice is the consistent style and personality a brand uses in its writing, covering word choice, rhythm, and attitude.

A set of related pages organized around one central topic, with a main pillar page linked to several supporting articles.

Topical authority is the depth and breadth of trustworthy content a site has on a subject, which helps it rank across an entire topic.

A transformer is a neural network architecture that uses a mechanism called attention to weigh the relationships between all parts of an input at once.

A vector database is a database built to store and search high-dimensional vectors, returning the items whose vectors are most similar to a query vector.

A software program that uses AI to hold a conversation with a person and complete the task they are asking for, rather than just answering a single question.

The structured practice of capturing what customers say about their needs, expectations, and experience, then turning that feedback into action.

Workflow automation is the use of software to run a defined sequence of tasks automatically, triggered by an event and following set rules without manual steps.

An XML sitemap is a structured file that lists the URLs on a site and helps search engines discover and crawl them efficiently.

Zero-shot learning is when a model performs a task it was never explicitly trained on, using only a description of the task and no labeled examples.