The Chatbot 101 Glossary

From the makers of AI Chatbots.

Zoe Ang
Zoe Ang
Nov 1 · 5 min read

Jargons are present in every industry to help industry players to communicate with one other efficiently. However, they tend to confuse the outsiders and lead to poor communication. Here is a list of jargons used in the “Chatbot World” and we hope this will help you navigate in the land of chatbot/technology like a pro.

Artificial Intelligence (AI)

Intelligence exhibited by machine. In computer science, an ideal “intelligence” machine is a flexible rational agent that perceives its environment and takes actions that maximize its chance of success at some goal.

Application Programming Interface (API)

The messenger that takes requests and tells a system what you want to do and then returns the response to you.

Bot Amnesia

A bot’s inability to maintain context in a conversation.

Bot Invocation

UI consideration for how to wake up a bot and initiate a conversation.

Business Logic

The pre-defined premise to run the business. Your chatbot will have to ascertain key facts from customers before providing the right solutions.

Context

Chatbot’s understanding of the situation explained by the user.

Contextual Questions

Questions which refer to something earlier in the conversation and are ambiguous on their own.

Conversation Management

Managing complex, multi-intent conversations.

Corpus

A collection of written or spoken materials stored about a topic.

Dialogue Management

Manage the memory and context in a single conversation and across conversations to conduct natural, human-like back and forth conversation.

Entity

Entities are key variables. It is a specific object type that exists separately from other things and has a clear identity. Entity modifies the intent as it provides more specifics.

Entity Mapping

Mapping variables we want to collect from the user to create context.

Entity Recognition/ Entity Extraction

Identifying entities which exists in the utterances to facilitate more complex commands and analysis

Escalation

Distinguish what is important and processing them with higher priority. In some chatbots, the questions can also be routed to live agents for complex or tricky situations.

Explaining Possibilities

AI assistants are always limited to helping users with a specific set of tasks and should be able to tell a user what they can do. That includes coherently responding to requests that are out of scope.

Explicit Confirmation

The chatbot asks the user to clarify how it should help.

Fallback

The backup plan executed when the user’s intent does not match existing intents and flows.

Flow

The overall dialogue flow following the trigger.

Handoff

Break out of a conversation and allow a human agent to answer the query. Handoffs are useful when users are getting frustrated at not getting the answers they are looking for or when the chatbot is not trained for a complicated task.

Humanize

Giving a personality and human touch to the chatbot.

Implicit Confirmation

Implicit confirmation involves repeating details back to the user to reassure them that they were understood correctly. This also gives the user a chance to intervene if your assistant misunderstood.

Intention

The user’s intention. What the user actually wants out of the conversation?

Intent Mapping

Matching user’s statements (utterances) to the correct intents

Intent Recognition

Identify what the user’s intent, even if phrased unexpectedly.

Memory

The chatbot’s understanding of historical utterances which aids intent recognition and directs the user to the right dialogue flow.

Menu Cards

A listing of what the chatbot is trained for in the form of buttons and words. This helps to direct users into the right dialogue flow.

Machine Learning (ML)

Training a program using data and information available (answers). In chatbot terms, chatbots learn how to respond to the user by analysing human agent responses. Necessary for qualitative intent recognition

Natural Language Processing (NLP)

The translation of human language to one which computer system can comprehend and vice versa.

Natural Language Understanding (NLU)

Understanding intents and extracting key variables (entities) from a user’s inputs.

Persistent Menus

As the user may get lost in the conversation, cancel a conversation, or context-switch to another taste, you will need to think about giving your users a solid understanding of how to navigate the bot conversation.

Prediction

The ability to predict the right answer to a question, or an action to take at a particular time in the conversation.

Proof-of-concept (POC)

A beta stage of chatbot development, where the chatbot is functional when its inputs are artificially constrained.

Q&A Pairs/ Scripts/ Conversation Scripting

Facts, details or solutions to queries or requests.

Reinforcement Learning

Chatbot learns from user “corrections” overtime to improve the suitability of responses.

Rich Controls

Use of contextual-relevant elements such as buttons, images, emojis

Sentiment Analysis

Understand the mood of the conversation. Is the user happy? Upset?

Small Talks

Refers to the general category of intents usually used for greetings, acknowledgements, chitchats and insults.

A natural conversation begins and ends with greetings like “Hello”, and users make statements which require acknowledgementssuch as “that’s awesome!”. The chatbots will often receive other forms of small talks which are out-of-scope. Queries like “will you marry me?” are not uncommon, and we classify them as chit chat. Angered users may also lash out at the bot with insults.

Supervised Learning

Machine trained using data which are well “labelled” or classified.

Triggers

Keyword or status change that starts a series of actions.

Turing test

A test for an intelligent computer system. If the end-user is unable to differentiate if a human-agent or a computer program held the conversation, we say the computer system passes the “turing test”.

Unhappy Path

Users who refuse to provide the required information or wish to correct something they said earlier or interrupts conversational flows with chitchats.

Unsupervised learning

Machine learning with data which are not “labelled”.

User Acceptance Test (UAT)

Testing conducted on the beta stage chatbot. Includes real world requirement to uncover areas of improvements.

User Experience (UX)

The overall experience of the usage of a program. In particular, how intuitive or easy, or pleasing it is to use.

User Feedback

Ask for feedback to understand if the bot was efficient.

User Interface (UI)

How an end-user and the program interact.

Utterance

Every statement made by an end-user.


Originally published at .

KeyReply

Stories about chatbots, NLP, and superpowers for your team

Zoe Ang

Written by

Zoe Ang

Avid data enthusiast & content creator at KeyReply

KeyReply

KeyReply

Stories about chatbots, NLP, and superpowers for your team

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade