{"id":20204,"date":"2020-01-29T12:22:19","date_gmt":"2020-01-29T12:22:19","guid":{"rendered":"https:\/\/www.arimetrics.com\/glosario-digital\/chatbot"},"modified":"2026-09-18T09:56:29","modified_gmt":"2026-09-18T09:56:29","slug":"chatbot","status":"publish","type":"encyclopedia","link":"https:\/\/www.arimetrics.com\/en\/digital-glossary\/chatbot","title":{"rendered":"Chatbot"},"content":{"rendered":"<p><img decoding=\"async\" class=\"boxpad alignright wp-image-14044 size-full\" src=\"https:\/\/www.arimetrics.com\/wp-content\/uploads\/2020\/01\/chatbot-1.jpg\" alt=\"Chatbot\" width=\"300\" height=\"300\" srcset=\"https:\/\/www.arimetrics.com\/wp-content\/uploads\/2020\/01\/chatbot-1.jpg 300w, https:\/\/www.arimetrics.com\/wp-content\/uploads\/2020\/01\/chatbot-1-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/><strong>Definition:<\/strong><\/p>\n<p>A <strong>chatbot<\/strong> is a software system that maintains a conversational interaction with a person through text, voice or other compatible inputs and outputs.<\/p>\n<p>It may provide information, collect data, guide a process or perform actions within a defined scope. <strong>Chatbot describes a conversational function and interface, not a particular technique:<\/strong> some systems follow rules, others retrieve information and others generate responses using artificial intelligence models.<\/p>\n\n<h2>How a chatbot works<\/h2>\n<p>A chatbot receives an input, relates it to the state of the conversation and determines a response or action. How it does this depends on its architecture and the systems with which it is integrated.<\/p>\n<p>An implementation may contain:<\/p>\n<ul>\n<li><strong>Channel and interface:<\/strong> A website, application, messaging service, telephone system or other environment used to exchange messages.<\/li>\n<li><strong>Input processing:<\/strong> Rules, intent classification, speech recognition or models that interpret the request.<\/li>\n<li><strong>Conversation state:<\/strong> Information required to maintain dialogue continuity, such as the current step, earlier responses or outstanding data.<\/li>\n<li><strong>Information sources:<\/strong> Approved content, knowledge bases, catalogues, internal systems or documents consulted to prepare a response.<\/li>\n<li><strong>Orchestration and actions:<\/strong> Logic that decides which response to provide, which tool to consult or which operation to request.<\/li>\n<li><strong>Output and transfer:<\/strong> The response presented to the person, confirmation of a result or transfer to another channel or human operator.<\/li>\n<\/ul>\n<p><strong>Conversation context is not necessarily equivalent to persistent memory.<\/strong> One system may retain information during a session and remove it afterwards, while another may associate later conversations with an identity or history. This difference needs to be explained and configured according to the service&#8217;s purpose.<\/p>\n<p>When a chatbot can make bookings, modify data, send messages or perform other operations, it requires proportionate permissions and validation. A generated response and an executed action are different outcomes.<\/p>\n<h2>Chatbot classification<\/h2>\n<p>A rule-based chatbot follows predefined options, keywords or paths. It is suitable for predictable processes but may fail when an input does not match the conditions that have been anticipated.<\/p>\n<p>A retrieval-based system selects responses or passages from a defined source. A generative chatbot uses a model to produce a new response according to its instructions, context and available information. Hybrid systems combine these methods and may apply different rules according to the task or level of risk.<\/p>\n<p><a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/chat\">Chat<\/a> is the space or conversation in which messages are exchanged. Live chat is commonly handled by a person, although one interface may combine automation and human support. The chatbot is the automated participant, not the entire channel.<\/p>\n<p>The expression <strong>virtual assistant<\/strong> may refer to a chatbot with access to several functions, but it can also describe systems that do not depend primarily on conversation. An <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/ai-agent\">AI agent<\/a> may plan and carry out multistep processes using tools. A chatbot may incorporate agent capabilities, but the two terms are not equivalent.<\/p>\n<p>A voicebot adds speech recognition and synthesis to the interaction. Voice changes the input and output mode but does not by itself determine how responses are selected.<\/p>\n<h2>Chatbot applications<\/h2>\n<p>Chatbots are used in customer service, ecommerce, bookings, internal support, training and access to knowledge bases. They may answer frequently asked questions, check a status, collect initial information or support a task.<\/p>\n<p><strong>Each implementation needs to define what the chatbot can do and what remains outside its scope.<\/strong> The system needs to recognise requests it does not understand, matters requiring a person and situations where it lacks sufficient information.<\/p>\n<p>Its automated identity needs to be understandable where confusion about the nature of the interlocutor could arise. It is also useful to explain which information is requested, whether history is retained and how human support can be reached when available.<\/p>\n<p>Transfer needs to provide the necessary context without requiring the entire conversation to be repeated, while avoiding the disclosure of irrelevant data. Staff availability, working hours, queues and response times remain part of the service even when initial access to the chatbot is continuous.<\/p>\n<p>Design also needs to consider clear language, keyboard navigation, alternatives to voice, understandable error messages and mechanisms for reviewing or cancelling an action.<\/p>\n<h2>Chatbot governance<\/h2>\n<p>Evaluation needs to correspond to the task. Resolved requests, correct responses, completed operations, errors, abandonment, transfers, repeated enquiries, response time and satisfaction may be analysed. A high containment rate is not beneficial if it prevents access to a person or prolongs unresolved conversations.<\/p>\n<p>Generative systems can produce incorrect or unsupported claims. The <a href=\"https:\/\/www.nist.gov\/publications\/artificial-intelligence-risk-management-framework-generative-artificial-intelligence\" target=\"_blank\" rel=\"noopener\">NIST Generative AI Risk Management Profile<\/a> includes these confabulations among the risks that need to be identified and managed. Testing needs to cover normal, ambiguous, incomplete and adversarial requests.<\/p>\n<p>When a chatbot uses external documents, pages or tools, an input may attempt to alter its instructions. <a href=\"https:\/\/genai.owasp.org\/llmrisk\/llm01-prompt-injection\/\" target=\"_blank\" rel=\"noopener\">OWASP identifies prompt injection<\/a> as a particular risk for language-model applications. Restricting permissions, separating data from instructions, validating results and requiring confirmation before relevant actions can reduce the potential impact but do not guarantee that failures will not occur.<\/p>\n<p>Conversations may contain personal data, credentials or confidential information. Purpose, access, retention, deletion and the providers involved need to be defined. <strong>Automation does not remove the responsibility to supervise the service, maintain its sources and address the effects of incorrect responses or actions.<\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Definition: A chatbot is a software system that maintains a conversational interaction with a person through text, voice or other compatible inputs and outputs. It may provide information, collect data, guide a process or perform actions within a defined scope. Chatbot describes a conversational function and interface, not a particular technique: some systems follow rules, [&hellip;]<\/p>\n","protected":false},"author":6,"featured_media":79984,"template":"","encyclopedia-tag":[1385,1379],"class_list":["post-20204","encyclopedia","type-encyclopedia","status-publish","has-post-thumbnail","hentry","encyclopedia-tag-ai-agents","encyclopedia-tag-digital-communication"],"_links":{"self":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia\/20204","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia"}],"about":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/types\/encyclopedia"}],"author":[{"embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/users\/6"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/media\/79984"}],"wp:attachment":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/media?parent=20204"}],"wp:term":[{"taxonomy":"encyclopedia-tag","embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia-tag?post=20204"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}