Mixed Deck — All Chatbots Topics Flashcards

What is a 'transactional chatbot' in the context of e-commerce?

Answer: A bot that can complete actions like placing orders, processing returns, or making payments within the conversation

Transactional chatbots go beyond Q&A to execute real business actions, creating end-to-end commerce experiences within the chat interface.

What does a chatbot 'fallback' response do?

Answer: Provides a default reply when the bot doesn't understand input

A fallback gives a graceful default message when no intent is matched.

What is 'error recovery' in chatbot design?

Answer: Providing helpful guidance when a user's input cannot be understood

Effective error recovery acknowledges misunderstanding and offers clarifying questions or alternative paths.

What metric measures how often a chatbot resolves issues without human help?

Answer: Containment or self-service rate

Containment rate tracks the share of conversations fully handled by the bot.

What is a 'multi-turn conversation' in chatbots?

Answer: A dialogue with several back-and-forth exchanges

A multi-turn conversation involves multiple exchanges between user and bot.

What are chatbot communication forms?

Answer: Both options are correct

Chatbots communicate with users through both text and voice. Text-based chatbots are common on websites and messaging applications, while voice-based chatbots, like virtual assistants, interact using spoken language. Therefore, both text and voice are correct forms of communication for chatbots.

Which practice improves chatbot NLP model performance over time in production?

Answer: Continuously retraining the model with new real-world conversation data and corrected intent labels

Ongoing retraining with production data keeps the model current with evolving user language patterns and new topics.

What is a 'webhook' in the context of chatbot development?

Answer: An HTTP callback that delivers real-time event data from a messaging platform to the bot's backend

Webhooks allow messaging platforms like Facebook Messenger or Slack to push user messages to the bot server in real time.

What is the role of a 'language model' in modern chatbots?

Answer: To predict and generate contextually appropriate text based on learned patterns from large corpora

Language models provide the core ability to understand and generate human-like text that powers modern conversational AI.

What does a 'conversation flow' define in a chatbot?

Answer: The structured path a dialogue follows

A conversation flow maps out the structured path and steps a dialogue can take.

What is a 'fallback response' in a chatbot?

Answer: A default reply when the bot doesn't understand input

A fallback response is the default message given when the bot cannot understand the user.

What is the purpose of a 'confidence score' in intent matching?

Answer: To indicate how certain the bot is about the matched intent

A confidence score shows how sure the model is that it matched the right intent.

Which integration method allows a chatbot to look up real-time data like inventory or account balances?

Answer: Live API calls to backend systems during the conversation

Real-time API integrations let chatbots fetch up-to-date information from backend systems to give accurate, live responses.

A 'quick reply' button in a chat interface is designed to:

Answer: Offer preset response options the user can tap

Quick replies give tappable preset options to speed up interaction.

What is 'fine-tuning' an AI chatbot model?

Answer: Further training a model on specialized data for a task

Fine-tuning adapts a pretrained model to a specific domain or use case.

What role does audit logging play in chatbot security?

Answer: It creates a tamper-evident record of all chatbot interactions and system events, enabling forensic analysis after a security incident

Audit logs provide an immutable record of who did what and when, which is essential for investigating breaches, meeting compliance requirements, and detecting anomalous behavior.

What role do 'typing indicators' play in chatbot UX?

Answer: They simulate human-like processing time, reducing the perception of lag

Typing indicators create a natural conversational rhythm and signal to users that the bot is working on a response.

Joseph Weizenbaum defined ELIZA as the first _______-based chat program.

Answer: NLP

Joseph Weizenbaum defined ELIZA as the first Natural Language Processing (NLP)-based chat program. Although its NLP capabilities were rudimentary by today's standards, ELIZA was groundbreaking in its ability to process and respond to human language in a conversational manner. It demonstrated the potential for computers to interact using natural language, a core goal of NLP.

Which evaluation metric measures how well a chatbot's NLP model avoids false intent matches?

Answer: Precision — the proportion of predicted intents that are actually correct

Precision ensures the model doesn't incorrectly assign an intent when the user actually meant something else.

Why is 'data augmentation' important in NLP chatbot training?

Answer: To artificially expand training datasets with paraphrased examples, improving model robustness

Augmenting training data with paraphrases and variations helps the model generalize better across diverse user phrasings.