The words behind the work.

A software system that can use information and tools to complete a bounded task. Its permissions, checks, and handoffs matter more than whether it uses a chat interface.

Computer systems that perform tasks associated with learning, reasoning, perception, or language. The useful business question is which task they perform reliably and how results are checked.

A repeatable process carried out by software with limited manual effort. Many valuable automations use fixed rules and do not need a language model.

Assigning an item to a defined category, such as routing an incoming support request by topic or urgency.

A sequence that collects, cleans, transforms, and delivers data to another system. For example, vendor documents can become structured purchase records.

How complete, current, consistent, and fit for purpose data is. A model cannot repair an unreliable source of truth by itself.

A numerical representation of content used to compare meaning. Embeddings can help find relevant documents even when a question uses different words.

Additional training of a model on task-specific examples. It is one option for changing behavior, not a substitute for current source data, evaluation, or access controls.

An answer based on identified source material that a person can inspect. A citation helps review, but the source and the answer still need checking.

A confident-sounding output that is unsupported or false. Source retrieval, constrained tasks, testing, and human review can reduce the risk.

A workflow in which a person reviews or approves consequential steps, such as changing an order or resolving a member dispute.

An organized set of approved documents, policies, or answers that staff and assistants can search and maintain.

A model trained to work with language and other content. It can summarize or draft, but its output needs boundaries and evaluation for the job at hand.

Able to process or produce more than one kind of information, such as text, images, audio, or video.

Turning text in scans or images into machine-readable text so it can be searched, extracted, or reviewed.

The instructions and context given to a model for a task. A good prompt does not replace reliable data, permissions, or testing.

A pattern that retrieves relevant material at answer time and gives it to a model as context. It helps use current sources without retraining the model.

A model’s request to use a defined function or system, such as looking up an order. The application still controls access, validation, and execution.

Searching numerical representations of content to find similar meaning rather than only exact keywords.

The steps, people, systems, and decisions that move a task from trigger to result. Mapping it is usually the first step before choosing automation.

Designing and building a service so people with different abilities and assistive technologies can use it. It includes content, interaction, code, and testing.

Text that conveys an image’s information or function to people who cannot see it. Decorative images generally use an empty alt attribute.

An application programming interface: a defined way for software systems to request data or actions from each other.

A content management system lets editors create and update site content without changing application code. It still needs clear publishing roles and content structure.

Google’s user-experience measures for loading, responsiveness, and visual stability. They help identify real page experience problems rather than a single generic “speed score.”

A customer relationship management system holds records of prospects, customers, interactions, and sales activity so teams can coordinate follow-up.

An open-source content management system used to build and manage websites and digital experiences.

Enterprise resource planning software that connects operational records such as purchasing, inventory, production, and finance.

Extract, transform, load: moving data from source systems, cleaning it, and delivering it to a destination for reporting or operations.

The markup that gives a web page its structure and meaning. Semantic HTML helps browsers and assistive technologies understand content.

Integration platform as a service: hosted tooling for connecting applications and moving data between them.

A software architecture in which multiple customers use a shared application while their data and permissions remain separated.

An authorization framework that lets an application receive limited access to another service without receiving the user’s password.

Layouts and interactions that adapt to different screen sizes and input methods without hiding essential content.

Software as a service: software delivered over a network and typically managed by its provider rather than installed and maintained by each customer.

The address of a resource on the web. A stable, descriptive URL helps people share and return to the right page.

A service exposed over a network so other software can exchange data or request an action through a defined interface.

A message one system sends to another when an event occurs, such as a form submission or payment update.

An HTTP response that permanently points a URL to a new location. Use it when an old page has a close replacement.

An HTTP response that temporarily points a URL somewhere else. Use it when the move is not intended to be permanent.

Structuring clear, well-supported answers so search and AI answer features can understand and reference them. It builds on useful content and sound SEO.

The preferred URL for a page when similar or duplicate versions exist. A canonical hint helps search engines consolidate those versions.

Whether a search engine can discover and fetch a page through links and permitted requests.

Substantially similar content at more than one URL. It is usually a URL-selection and consolidation problem, not proof of a penalty.

Improving how a brand’s useful content is understood and potentially cited in generative search experiences. No markup or tactic guarantees inclusion.

A search engine’s process of understanding and storing a page for possible use in search results. Being crawlable does not guarantee indexing.

A link from one page on a site to another page on that same site. Useful links help people and crawlers reach related information.

A word or phrase people use to search. The goal is to answer the underlying question, not repeat the phrase a set number of times.

A page designed around a specific visitor need or campaign action, with a clear next step and enough information to make a decision.

A short HTML description of a page. Search engines may use it as a search-result snippet, but may generate a different snippet when that better matches a query.

A search result shown because the search engine considers it relevant, rather than because the site paid for an ad placement.

The page of results and search features returned for a query.

The goal behind a search, such as learning a concept, comparing options, or finding a specific service.

Search engine optimization: making a site useful, accessible, and understandable so relevant pages can be discovered and shown in search.

Machine-readable markup that describes entities or page content using a defined vocabulary, such as schema.org. It supplements visible content; it does not replace it.

The site-level work that helps search engines fetch, render, understand, and index the right URLs, including structure, performance, redirects, and metadata.

A file listing URLs a site wants search engines to discover. It is a discovery hint, not a guarantee that every listed page will be indexed.

The method used to assign credit for a lead or sale across marketing interactions. Every model makes assumptions, so report the model alongside the result.

A clear invitation to take the next useful step, such as requesting an evaluation or talking with a specialist.

Invalid or deceptive ad clicks that can distort campaign data and spend. Detection and remediation depend on the ad platform and evidence available.

Clicks divided by impressions, expressed as a percentage. It describes response to a listing or ad, not the quality of the resulting visits.

A visitor action a business has chosen to measure as valuable, such as a qualified form submission or purchase.

Conversions divided by the relevant opportunities to convert, such as sessions or users. State the denominator when comparing rates.

Ad spend divided by ad clicks. It is one cost measure, not evidence that the clicks became qualified leads.

Information an organization collects directly through its own interactions with people, subject to the applicable notice and consent choices.

Google Analytics 4, an event-based analytics platform for measuring site and app interactions. Events need clear definitions and privacy controls.

A count of when a listing, ad, or other item is shown under a platform’s measurement rules. Definitions can differ by platform.

A way to rank leads against agreed signals, such as fit and activity. Scores should be checked against actual sales outcomes and revisited over time.

A lead that meets a team’s documented marketing criteria for follow-up. The criteria should match the sales process, not a universal template.

A sequence of useful follow-ups tied to a person’s stated interests or actions, with clear timing and opt-out handling.

A paid advertising model in which an advertiser pays when someone clicks an ad.

Marketing through search engines, commonly used to mean paid search campaigns. Define the scope when comparing it with SEO.

Tags added to a URL to identify a campaign, source, or medium in analytics. Use consistent naming and avoid putting personal information in the URL.

A service that uses available signals to suggest which organizations may have visited a site. Coverage and confidence vary; a company match is not proof of who the individual visitor was.

Measurement of site use and outcomes, such as visits, engagement, qualified inquiries, and completed forms. Useful reporting begins with agreed goals and consent-aware collection.