Chatbots glossary, terms, and definitions

Terms and definitions

Agent: What is an agent?
A customer service agent is the person tasked with answering and assisting customers with their enquiries. Agents are usually organised within teams, each servicing a different area of expertise or communications channel.
Algorithm: What is an algorithm?
An algorithm is a sequence of instructions designed to solve a problem. Depending on their type, algorithms make use of different types of data to produce the desired results. They are widely used in mathematics and computer science to solve problems of varying degrees of complexity.
Artificial Intelligence: What is artificial intelligence?
Artificial intelligence is an umbrella term describing various computing techniques and algorithms designed to mimic human thought processes. Artificial intelligence is used whenever a computer program requires to make decisions based on ambiguous data.
AI-based chatbots: What are AI-based chatbots?
AI-based chatbots are chatbot programs designed to engage the user in human-like conversation in order to accomplish common tasks. They make use of advanced algorithms to deduce meaning and context from user statements, allowing them to provide more interactivity than a standard user interface.
Autoresponder: What is an autoresponder?
An autoresponder is a software program which automatically responds to incoming messages with a pre-set message. It can be a standalone application serving many different communication channels or a function within a larger messaging system.
Bot builder: What is a bot builder?
A bot builder is an application that is designed to help non-programmers create chatbots without the need for specialised training. These applications are developed by chatbot service providers in order to make their services as accessible as possible.
Broadcast: What is a broadcast?
A broadcast is a message sent to every user of a platform, service, or communication channel at once. Broadcasts can be used to disseminate messages from a brand to their customers.
Call to action: What is a call to action?
A call to action is usually the final element of a marketing message, intended to suggest a response from the end user. These can be presented as buttons, statement encouraging the user to read more et cetera.
Communication Channel: What is a communication channel?
A communication channel is any method of contact between the brand and its customers. They include traditional channels, such as mail and phone calls, as well as modern channels, such as social media.
Chatbot: What is a chatbot?
A chatbot is a piece of software designed to emulate human-to-human conversations. Chatbots are used to present a user-friendly interface through which customers can take care of various commonly requested tasks – check their bank account balance, order airplane tickets, track orders et cetera.
Chatbot persona: What is a chatbot persona?
A chatbot persona is a set of guidelines regarding a bot’s presentation. These rules include its name, gender (if it has any), associated imagery (such as avatars) and its tone of presentation. Voice chatbots (voicebots) also define a specific voice pattern, accent, and delivery.
Cloud computing: What is cloud computing?
Cloud computing refers to the practice of using dynamic, on-demand computing resources. Instead of maintaining their own expensive data centers, companies instead rent computer time from large server farms to host their services. This approach is more energy-efficient and economical than running your own server farm.
Conversational User Interface: What is a conversational user interface?
A conversational user interface allows users to communicate with a system the same way they communicate with people – by talking to it! These types of interfaces are divided into text-based interfaces and voice-based interfaces.
Customer persona: What is a customer persona?
A customer persona is an imagined depiction of an average person belonging to a target customer group. This persona is composed of insights drawn from market research as well as assumptions about the ideal target group. Customer personas are extremely useful in creating marketing and customer service strategies.
Customer satisfaction score: What is a customer satisfaction score?
A customer satisfaction score (commonly abbreviated as CSAT) is a metric used to express a given customer’s satisfaction with a brand. The data for this metric is drawn from surveys presented to customers, asking them to rate their satisfaction on a linear scale.
Deep Learning: What is deep learning?
Deep learning is a subset of machine learning techniques based on neural networks. These algorithms emulate the way a human brain works in order to provide a much higher level of flexibility and rate of learning than competing algorithms, at a cost of complexity and resources.
Deployment: What is a deployment?
Deployment is the term used to describe the process of launching a product or service. A deployed service is often called “production” in order to differentiate it from one in “testing” or “development”.
Entity: What is an entity?
An entity is any logical element of a statement – the topic, auxiliary data, or grammatical structures.
Entity extraction: What is entity extraction?
Entity extraction is the process of identifying discrete parts of any given statement. This is used to determine the statement’s subject, it’s emotional sentiment, as well as additional information that can be useful in the overall process.
Flow-based chatbot (Rule-based chatbot): What is a flow-based chatbot?
A flow-based chatbot, also known as a rule-based chatbot works using a predetermined dialogue flow. Unlike AI-based chatbots, it can only operate within the rigid structure it was programmed for. These chatbots are cheaper and easier to deploy, but their capabilities remain static; they don’t learn or self-improve.
Historical data: What is historical data?
Historical data is the record of every communication between your brand and its clients. This data can be used for statistical analysis purposes as well as a training dataset for automation solutions – such as chatbots.
Human fallback: What is human fallback?
Human fallback is the process of a chatbot handing off a case to a human agent. This happens when the customer’s request cannot be handled by a chatbot.
Intent: What is an intent?
An intent is the purpose for which the customer gets in touch with the brand. While numerous, they often can be categorised within larger, recurring groups, for which discrete customer service teams can be organised.
Intent detection: What is intent detection?
Intent detection is the process of identifying the customer’s intent from their messages. Natural language processing algorithms can determine a message’s intent with a high degree of accuracy based upon historical data.
Interface: What is an interface?
An interface is any method through which an user communicates with a system. These can be graphical, text-based, conversational, voice, and many more. Different types of interfaces offer different degrees of interaction within a given system.
Machine Learning: What is machine learning?
Machine learning is the process through which an algorithm learns and self-improves its behaviour based on provided training data. Machine learning is extremely versatile and has a broad range of applications, including, but not limited to image recognition, text analysis, self-driving vehicles, and more.
Natural Language Generation: What is natural language generation?
Natural language generation is the process of constructing grammatically-correct sentences through algorithms. Natural language generation is an important technique in building AI-based chatbots – it allows them to construct their own sentences rather than using predetermined statements.
Natural Language Processing: What is natural language processing?
Natural language processing is the term used to describe a broad range of algorithms designed to extract computer-workable data from statements made in a natural language, such as English.
Natural Language Understanding: What is natural language understanding?
Natural language understanding is the process of extracting information from natural language statements and putting that information in a workable context. This technique allows chatbots to easily adapt to new information over the course of a conversation.
Neural network: What is a neural network?
A neural network is computer software designed to emulate the behaviour of human neurons. Neural networks are used in machine learning to provide a degree of prediction and adaptability to the algorithm.
Omnichannel: What is omnichannel?
Omnichannel is a marketing and customer service strategy which involves the brand being proactively present and available for contact on every social media platform their customers might be on. It differs from multichannel, which only deals with a predetermined selection of channels.
Proof of concept: What is a proof of concept?
A proof of concept implementation is a preliminary version of a software product, containing only the necessary functionality. It is usually intended for demonstration purposes – for instance, to show off to potential investors or customers.
Task: What is a task?
A task is any action a bot can be programmed to perform. Usually, chatbots are programmed to take care of the tasks most commonly requested by customers, in order to lighten the workload placed on the customer service department.
Training dataset: What is a training dataset?
A training dataset is any collection of data used to train a machine learning algorithm. The program learns by taking in new information and comparing it with data it already knows in order to determine the similarity between the two. With time, the algorithm learns to accept a wide variety of inputs with a very high degree of accuracy.
Trigger: What is a trigger?
A trigger is any phrase that triggers a certain process or reaction from a chatbot. A commonly encountered trigger is the fallback request, which signals to the bot that the customer wishes their case to be handed off to a human agent.
Turing test: What is the Turing test?
The Turing test is a proposed experiment to test a machine’s ability to exhibit intelligent, humanlike behaviour. A machine is considered to pass the test if a human observer cannot tell the difference between it and a human performing the same task. The test is named for its inventor, Alan Turing.
Universal artificial intelligence: What is universal artificial intelligence?
Universal artificial intelligence is a theoretical approach to artificial intelligence which aims to mathematically express a “formula for intelligence”. Research into this theorem resulted in significant contributions to AI research, computer science, and more.
User experience: What is user experience?
User experience is the term used to describe the entirety of a user’s interaction with a system – and the user’s emotional reaction to it. User experience designers take utmost care to ensure that systems are intuitive and easy to comprehend.
Utterance: What is an utterance?
An utterance is a single statement made by a user. Utterances are analysed by chatbots in order to determine their intent. This helps them understand the context of the conversation and offer relevant answers.
Virtual agent: What is a virtual agent?
Virtual agents are chatbots designed for customer service purposes. They usually take care of the most menial and recurring tasks customers expect of a customer service department – from checking account balances to tracking packages.

Case studies

Use cases

EU Projects

Chatbots glossary

Table of contents

Terms and definitions