A glossary of terms related to AI, chatbots and conversational marketing. Learn the lingo and become a chatbot expert.
A feature that allows models to consider larger amounts of information.
The simulation of human intelligence processes by machines, especially computer systems.
A set of rules or procedures for solving a problem or accomplishing a task.
A machine learning approach where the model queries the user or an oracle for input on uncertain data.
Advanced AI systems capable of performing tasks in the real world.
Application Programming Interfaces that enable GPTs to connect with other services and data.
An API designed for building conversational AI models.
Language models enhanced with external knowledge sources, such as embeddings or vector databases, to improve performance in tasks like retrieval and question answering.
Information or insights that can be directly applied to solve problems, take actions, or inform decisions in AI systems.
A workflow structure in which AI agents autonomously plan, reason, and execute sequences of tasks using available tools and knowledge.
An AI system capable of making decisions, planning, and acting independently to achieve specific goals.
A mechanism for controlling how frequently users or agents can make API requests to prevent abuse and ensure fairness.
The ability of an agent or LLM to store, recall, and leverage past interactions or information during task execution.
The process of improving a chatbot's performance by feeding it data and refining its algorithms.
The character or persona a chatbot exhibits during interactions, defined by its tone, style, and manner of communication.
An encoder that independently generates embeddings for queries and documents for scalable retrieval.
User interfaces that enable interaction with the user in a conversational manner.
The ability of a system to understand the context in which a user's input is given.
A platform or set of tools used to build, test, and deploy chatbots.
The sequence and structure of interactions that a chatbot follows during a conversation.
The process of integrating and making a chatbot live on a platform or channel for users to interact with.
An AI system's ability to constantly adapt and improve its performance by learning from new data over time.
A record of past interactions and conversations between the user and the chatbot.
A software program designed to simulate conversation with human users.
The process of crafting effective and natural dialogues for chatbots and virtual assistants.
Quantitative measures used to evaluate and optimize the performance, efficiency, and user satisfaction of chatbots.
A software or service that provides tools and infrastructure to design, develop, train, and deploy chatbots.
Features allowing users to set preferences for how ChatGPT operates.
Tailored versions of ChatGPT for specific tasks or knowledge areas.
A subscription level for users to access advanced GPT models.
A retrieval technique that incorporates the broader context of a query, such as user history or dialogue flow.
A generation technique where AI models create content based on the broader context of the input, including user history and prior interactions.
A type of encoder that jointly processes query-document pairs to determine relevance.
The range of input tokens or text that an AI model considers when generating or retrieving responses.
Embeddings that capture the meaning of words or phrases based on the surrounding context.
The technique of dynamically adding relevant context or data into LLM prompts or agent workflows to improve accuracy and relevance.
A prompting technique where LLMs generate multi-step reasoning, explanations, or solutions by thinking step by step.
A system or tool that allows LLMs to write, execute, and debug code in real-time as part of reasoning or automation.
A type of machine learning that uses neural networks with many layers.
The component of a chatbot that handles the flow of conversation.
The process where a chatbot seeks clarity on ambiguous user input to provide the most accurate response.
The latest version of OpenAI’s image generation model.
The process of retrieving relevant documents from a collection, often powered by vector embeddings and semantic search.
A retrieval method that uses dense vector embeddings, enabling semantic search and advanced contextual retrieval.
The process of organizing and storing documents in a structured format to enable efficient retrieval.
A measure of how similar two or more documents are based on their content or embeddings.
Specific pieces of information extracted from user input, like names, dates, and products.
The process of identifying and classifying key information or data points from user input in a chatbot conversation.
A training approach where a system learns to map inputs directly to outputs, minimizing intermediate steps or feature engineering.
Dense numerical representations of data, such as text or images, used in tasks like semantic search and retrieval.
Structured or unstructured data repositories used by AI systems to retrieve information and enhance responses.
The process of ensuring embeddings from different models or datasets are compatible for comparison or integration.
A default response given by a chatbot when it cannot understand or process the user's input.
A mechanism that allows systems to learn from their actions by receiving feedback on their performance.
Adjusting a pre-trained model to perform better on specific tasks or datasets.
AI’s ability to execute specific tasks or functions within a program.
The process of adapting retrieval models to specific tasks or datasets by training them on task-relevant examples.
An approach where AI models are trained to perform tasks with only a few labeled examples.
An interface that allows LLMs or agents to trigger external functions, APIs, or tools programmatically during reasoning.
A prompting method where models are shown a small number of examples to guide their outputs for new tasks.
A marketplace for sharing and discovering custom GPTs.
An advanced GPT-4 iteration with enhanced speed and performance.
AI systems designed to create new content, such as text, images, or audio, based on learned patterns and input context.
A training phase where AI models learn to predict and generate text based on large-scale datasets.
Policies, rules, or constraints that ensure AI models act safely, ethically, and within desired boundaries.
A model where human intervention assists in the decision-making process of an automated system.
A search method that combines dense and sparse retrieval techniques to improve accuracy and relevance.
A training method where human preferences or corrections are used to align and improve AI model behavior.
The ability of a chatbot to understand and identify what a user wants to achieve.
Set of protocols and tools that allow different software applications to communicate and work together.
The process of determining the specific goal or purpose behind a user's input in a conversation with a chatbot.
A method where models are guided to perform tasks using examples provided in the input prompt.
The process of updating retrieval system indices to reflect changes in the underlying data.
A format enabling structured data handling by AI models.
A lightweight, standardized protocol for remote procedure calls (RPC) using JSON, commonly used in AI system integrations.
A centralized repository of information that chatbots or systems use to answer user queries.
The process of extracting relevant information from a knowledge base or database, using methods like semantic search and embeddings.
A generative AI approach where outputs are grounded in external knowledge sources, such as documents or databases.
A technique where a smaller model learns from a larger, more complex model, retaining critical knowledge while reducing size.
Structured representations of information, linking entities and their relationships to facilitate efficient knowledge retrieval.
The process of incorporating external knowledge into AI models to improve performance and accuracy.
The date or point in time after which an AI model does not have knowledge of new events or data.
A technique that enhances AI model outputs by integrating retrieved knowledge into the generation process.
A real-time communication method between customers and support agents or sales representatives via a website or application.
A standardized markdown file located at the root of a website, designed to provide large language models (LLMs) with concise, structured information about the site’s content.
The process of coordinating large language models (LLMs), tools, and workflows to accomplish complex tasks.
A subset of AI that allows systems to learn and improve from experience without being explicitly programmed.
Interactions involving multiple modes or channels of communication, such as voice, text, and visuals.
Software layers that offer services and capabilities between the chatbot platform and external systems or databases.
The process by which machine learning models learn from data.
The ability of AI to understand and generate different data types like text and images.
A method where external memory modules or databases are integrated with AI systems to enhance their knowledge and context retention.
A conversational AI approach where retrieval systems provide relevant information for multi-turn interactions.
Servers implementing the Model Context Protocol (MCP) to enable advanced, tool-augmented, and multi-modal LLM interactions.
A software component that communicates with MCP (Model Context Protocol) servers to enable advanced AI orchestration and tool-augmented workflows.
A branch of AI that deals with the interaction between computers and human language.
Computing systems inspired by the structure and functioning of the human brain.
A training technique where irrelevant data points are sampled to improve the performance of retrieval models.
A retrieval method that uses deep learning models to generate embeddings and match queries with documents.
A systems challenge where every one of N tools must be integrated with every one of M models or agents, leading to exponential complexity.
A multi-channel approach to sales or customer service that provides an integrated and cohesive customer experience.
A task where AI systems answer questions using information retrieved from a wide range of unstructured data sources.
A core pattern in agentic AI where an agent observes the environment, reasons, and acts repeatedly to accomplish tasks.
A specific piece of data or instruction sent by a chatbot or received by it to trigger a certain action.
A technique where chatbots recognize specific patterns in user input to generate responses.
Machine learning models that have been previously trained on large datasets and can be fine-tuned or adapted for specific tasks.
Options for users to manage their data and how it's used by AI models.
A retrieval technique that identifies specific passages within documents to answer user queries effectively.
The practice of designing effective prompts to guide AI models in generating desired responses.
AI models that are pre-trained on large datasets to understand and generate human language effectively.
A modular system allowing developers to extend AI agent or LLM functionality by adding third-party tools, APIs, or workflows.
An AI application that provides accurate and context-aware answers to user queries based on a knowledge base or retrieved data.
The process by which AI systems produce replies or actions in response to user input.
A type of machine learning where models learn to make decisions through trial and error.
The consistency of AI in generating the same results under the same conditions.
A framework that combines retrieval and generation to produce responses grounded in external knowledge, leveraging techniques like dense retrieval and knowledge retrieval.
AI models that rely on retrieving relevant information, often using techniques like sparse retrieval or dense retrieval, rather than generating responses from scratch.
A framework combining retrieval and generation models to produce accurate, context-rich responses.
A model component used to encode queries and documents into embeddings for retrieval tasks.
The time it takes for a retrieval system to fetch relevant information in response to a query.
A system that combines retrieval and generation processes to enhance AI model outputs with relevant knowledge.
A technique that combines results from multiple retrieval methods to improve relevance and accuracy.
A tokenization method optimized for retrieval-augmented generation to balance efficiency and accuracy.
A system or component that enables AI agents to perform multi-step reasoning, logic, and decision-making.
A single interaction or series of interactions between a user and a chatbot during a specific timeframe.
A technique used to determine the sentiment or emotion expressed in a piece of text.
A technique used in conversational AI to gather specific pieces of information from the user.
Pre-defined replies or messages that a chatbot uses in specific scenarios or for certain user inputs.
A technology that converts spoken language into written text.
Measures to prevent misuse of GPTs and ensure ethical usage.
A retrieval method that uses traditional term-matching techniques, such as TF-IDF or BM25, to find relevant documents.
A search method that uses embeddings to understand the meaning behind queries and documents, enhancing retrieval relevance.
An instruction or set of guidelines given to an AI model to define its behavior, tone, or constraints for a session.
Data used to teach and refine machine learning models.
A test of a machine's ability to exhibit intelligent behavior indistinguishable from that of a human.
A predefined limit or value that determines specific actions or outcomes based on comparison with incoming data.
A technology that converts written text into audible speech.
A machine learning architecture used primarily in the field of natural language processing (NLP).
Technology that converts digital text into spoken voice output.
Vector representations of individual tokens, such as words or subwords, used in language models.
A technique or model that augments large language models with autonomous tool usage for enhanced reasoning and real-world actions.
The ability of AI agents or models to utilize external tools, APIs, or systems as part of their reasoning or action process.
A design approach for integrating autonomous tool use and decision-making within large language models.
Any input or phrase that a user communicates to a chatbot during a conversation.
A new iteration of GPT-3.5 with expanded capabilities and improved efficiency.
A metaphor for universal, plug-and-play standards in AI—enabling effortless compatibility between models, tools, and platforms.
A digital assistant that uses voice recognition to interpret and respond to user commands.
Databases designed to store and query high-dimensional vector embeddings for tasks like semantic search and dense retrieval.
Automated messages sent from apps when something happens.
A small software component that provides a specific functionality, often used to integrate chatbots into websites or apps.
A machine learning approach where models perform tasks without having seen labeled examples for those tasks during training.
Join over 1000+ businesses, websites and startups automating their customer support and other tasks with a custom trained AI agent.