The AI vocabulary can feel impenetrable — a mix of computer science jargon, marketing buzzwords, and genuinely useful concepts that travel professionals need to understand. This glossary cuts through the confusion, defining 45 AI terms with clear explanations and travel-specific context.
Terms are organised alphabetically.
A
Adaptive Learning
An AI approach where training content adjusts to each learner's demonstrated knowledge level. AI-powered adaptive platforms assess what an agent knows, skip what they've mastered, and focus on their gaps. An agent who knows Mediterranean hotels well gets advanced selling scenarios instead of introductory destination content. The opposite of one-size-fits-all training.
Algorithm
A set of rules or instructions that a computer follows to solve a problem or make a decision. In travel, algorithms power everything from revenue management pricing to the personalised learning paths in AI training platforms. Not all algorithms are AI — a simple sorting algorithm isn't intelligent. AI algorithms learn and improve from data.
API (Application Programming Interface)
A set of protocols that allows different software systems to communicate. In travel, APIs connect booking systems, CRMs, training platforms, and other tools — enabling data to flow between systems. You don't need to understand API technical details, but knowing the term helps when evaluating AI vendors.
Artificial Intelligence (AI)
The broad field of computer science focused on creating systems that perform tasks requiring human intelligence — learning, reasoning, problem-solving, and understanding language. In travel, AI encompasses training platforms, coaching engines, recommendation systems, revenue management, and content generation.
Automated Decision-Making
When AI systems make decisions without human intervention — setting a hotel room price, prioritising a customer enquiry, or recommending a training module. GDPR gives individuals rights regarding automated decisions that significantly affect them, which is relevant for AI in agent performance management.
B
Bias (Algorithmic)
Systematic errors in AI outputs caused by biased training data or flawed model design. In travel, bias might cause AI coaching to favour certain selling styles, recommendation engines to over-represent popular destinations, or pricing systems to charge differently based on customer location. Ethical AI practice requires monitoring and correcting for bias.
Big Data
Datasets too large or complex for traditional data processing. In travel, big data includes millions of booking records, social media mentions, search queries, and customer interactions. AI transforms big data from an overwhelming volume of information into actionable insights.
C
Chatbot
A software application that simulates conversation with humans, typically for customer service. Travel chatbots handle routine enquiries (booking status, baggage policies, visa requirements). Modern chatbots use NLP; older ones use rule-based responses. Often the most visible — but not the most valuable — AI application in travel.
Cloud Computing
Delivering computing services (storage, processing, AI) over the internet rather than on local computers. Cloud-based AI training platforms require no local installation and are accessible from any device — essential for mobile learning and distributed teams.
Computer Vision
AI that interprets visual information — images and videos. Travel applications include property image categorisation, visual search for destinations, and document processing (scanning passports and booking confirmations).
Conversational AI
AI systems that engage in natural dialogue — going beyond chatbots to understand context, maintain conversation history, and respond to nuanced queries. AI roleplay simulations are a form of conversational AI, creating realistic virtual customers for agent practice.
D
Data Mining
Extracting useful patterns from large datasets. In travel, data mining might reveal that agents who complete Caribbean product training within 7 days of a new product launch convert 40% more Caribbean enquiries — an insight that informs training timing strategies.
Deep Learning
A subset of machine learning using multi-layered neural networks to process complex data. Powers the most sophisticated AI applications — natural language understanding, image recognition, and speech processing. The technology behind modern AI coaching systems that analyse conversation quality.
Dynamic Pricing
Using algorithms to adjust prices in real time based on demand, competition, and other factors. The most established AI application in travel — airlines and hotels have used dynamic pricing for over a decade. STR data shows 5-12% revenue improvements.
F
Feature Engineering
The process of selecting and transforming data features that an AI model uses for learning. In travel performance analytics, this means choosing which training metrics (completion rates, assessment scores, roleplay frequency) are most predictive of sales performance.
G
Gamification
Applying game elements (points, badges, leaderboards, streaks) to non-game contexts. In travel training, gamification in AI assessment platforms can increase engagement — but only when it rewards genuine learning, not speed. See also: microlearning.
Generative AI
AI systems that create new content — text, images, code, audio — rather than just analysing existing content. In travel, generative AI produces training modules, marketing copy, customer proposals, and roleplay scenarios. ChatGPT is the most widely known generative AI tool.
GDPR (General Data Protection Regulation)
The EU/UK regulation governing personal data processing. Highly relevant to AI in travel because training platforms, coaching systems, and recommendation engines all process personal data. Ethical AI implementation requires GDPR compliance from day one.
H
Hallucination
When AI generates plausible but factually incorrect information — stating a hotel has a feature it doesn't, or inventing visa requirements that don't exist. A significant risk when using general-purpose AI tools like ChatGPT for travel-specific information. Purpose-built platforms reduce hallucination by training on verified data.
Hybrid AI
Combining different AI approaches — for example, rule-based systems for compliance questions (where accuracy is critical) and machine learning for personalised recommendations (where flexibility is valuable). Most effective AI implementations in travel use hybrid approaches.
L
Large Language Model (LLM)
The AI architecture behind tools like ChatGPT, Claude, and Gemini — trained on vast text datasets to understand and generate natural language. LLMs power AI coaching feedback, content generation, and conversational roleplay in travel applications.
Latent Demand
Demand that exists but hasn't been expressed — customers who would book a particular product if they knew about it or were recommended it. AI recommendation engines help identify latent demand by suggesting products that match customer preferences but weren't part of their original search.
M
Machine Learning (ML)
A subset of AI where systems improve their performance by learning from data rather than being explicitly programmed. In travel, ML powers adaptive training (learning what each agent needs), pricing optimisation (learning demand patterns), and customer matching (learning which products suit which customers).
Microlearning
Training delivered in short sessions (typically 3-7 minutes). Not technically AI, but AI enables effective microlearning through adaptive content delivery that makes each short session maximally relevant. The format that achieves the highest engagement and completion rates in travel agent training.
Model (AI Model)
The mathematical framework that an AI system uses to make predictions or decisions. Like a recipe: the model defines how inputs (agent assessment responses) are processed to produce outputs (personalised learning pathway). Different models suit different tasks — classification models for sorting, generative models for content creation, recommendation models for matching.
N
Natural Language Processing (NLP)
AI that understands, interprets, and generates human language. Powers chatbots, AI coaching analysis, sentiment analysis, and content generation. NLP enables AI to understand not just what words mean individually, but what they mean in context — crucial for analysing selling conversations.
Neural Network
A computing architecture inspired by the human brain's structure — layers of interconnected processing nodes. Deep neural networks (with many layers) power the most capable modern AI. You don't need to understand the architecture; you need to know that neural networks enable AI to learn complex patterns that simpler approaches can't handle.
O
Overfitting
When an AI model learns the training data too specifically, losing its ability to generalise to new situations. In travel training context: an assessment system that only works well for the specific hotels it was trained on but fails for new products. Good AI platforms are designed to avoid overfitting.
P
Personalisation Engine
The AI component that customises content, recommendations, or experiences for individual users. In travel training, the personalisation engine analyses each agent's knowledge profile and delivers tailored learning pathways. In customer-facing applications, it recommends products matching individual preferences.
Predictive Analytics
Using historical data and AI to forecast future outcomes. In travel: predicting which agents will improve fastest, which customers will rebook, which products will be popular next season, or which training interventions will have the highest ROI.
Prompt Engineering
Crafting effective instructions for generative AI systems. When using AI content generation tools, the quality of the prompt determines the quality of the output. A prompt like "Write a hotel description" produces generic results; "Write a 150-word selling guide for agents positioning Hotel Sunshine for active couples aged 30-45" produces useful content.
R
Recommendation Engine
An AI system that suggests products or content based on user data and behaviour. In travel: suggesting destinations to customers, training modules to agents, or coaching focus areas to managers. Uses collaborative filtering ("agents like you also found this useful") and content-based filtering ("based on your assessment gaps").
Reinforcement Learning
An ML approach where the system learns through trial and feedback — adjusting its approach based on outcomes. In travel AI coaching, reinforcement learning helps the system improve its feedback quality over time — learning which coaching suggestions lead to the greatest agent improvement.
ROI (Return on Investment)
The financial return generated relative to the cost of an investment. AI in travel consistently shows strong ROI: 300-700% for training platforms, 400-700% for coaching tools, 150-300% for revenue management. Calculated as: (Revenue Uplift - Investment Cost) / Investment Cost × 100.
S
Sentiment Analysis
AI that determines the emotional tone of text — positive, negative, or neutral. In travel: analysing customer reviews, social media mentions, agent feedback, and customer communications to detect satisfaction levels and emerging issues.
Spaced Repetition
A learning technique presenting information at progressively increasing intervals to optimise long-term retention. Not AI itself, but AI automates and personalises the scheduling. The most evidence-backed method for combating the forgetting curve in agent training.
Supervised Learning
An ML approach where the system learns from labelled examples. In travel coaching: training the AI on examples of "good" and "poor" selling conversations (labelled by experts) so it can evaluate new conversations. Most AI coaching systems use supervised learning as a foundation.
T
Training Data
The dataset used to teach an AI model. Quality training data is essential for accuracy. AI platforms trained on travel-specific data perform significantly better for travel applications than general-purpose AI trained on internet-wide data — because travel has domain-specific language, patterns, and requirements.
Transfer Learning
Using knowledge from one AI task to improve performance on a related task. In travel: an AI model trained on hotel product training can transfer learning to cruise product training — the selling principles and knowledge structures overlap even though the specific products differ.
U
Unsupervised Learning
An ML approach where the system identifies patterns in data without labelled examples. In travel: clustering customers into segments based on booking behaviour (without being told what the segments should be), or identifying product groupings based on feature similarity.
V
Voice AI
AI systems that understand and generate spoken language. Emerging in travel for: voice-controlled room features in hotels, phone-based customer service automation, and voice-activated agent tools that provide information during customer calls.
Using This Glossary
This glossary covers the AI terms most commonly encountered in travel business contexts. For sales enablement terminology, see our Sales Enablement Glossary. For training-specific terminology, see our Travel Agent Training Glossary.
For a comprehensive introduction to AI in travel, read our Complete Guide to AI in Travel and Tourism.
This article is part of our AI in Travel & Tourism series. Related reading:
