Glossary of terms

General Terms  

AI (Artificial Intelligence) Machines or software performing tasks that normally require human intelligence, like understanding language, recognising images or making predictions. 

ML (Machine Learning) A subset of AI where systems improve performance by learning from data rather than being explicitly programmed. 

Generative AI (Gen AI) AI that can create new content (text, images, video, audio) based on patterns it has learned from existing data. Examples include ChatGPT, DALL·E, MidJourney. 

Prompt A command or input given to an AI to generate output. 

Chatbot A conversational AI that interacts with users via text or voice.  

Natural Language Processing (NLP) The ability of AI to understand, interpret and generate human language. 

Deepfake AI-generated content (often video or audio) that can convincingly mimic real people or events. 

Automation Using AI or software to perform repetitive tasks like for example ad placement or social posting. 

Adaptive AI AI capable of learning from outcomes and modifying its behaviour to achieve a goal more effectively. Unlike automation, adaptive systems improve continuously in response to changing data or environments. Often referred to as goal-directed adaptive behaviour. 

Goal-Directed Adaptive Behaviour A foundational definition of intelligence: systems that learn from the results of their actions to better achieve objectives over time. Adaptation, not imitation, is the true hallmark of intelligence. 

Task Automation The use of AI systems to free humans from repetitive or structured tasks (e.g., robotic process automation, chatbots, image tagging). 

Content Generation Using generative AI to produce text, visuals, video or music, accelerating creative output and production workflows. 

Human Representation AI Systems that simulate human-like behaviour or reasoning to make predictions or act as stand-ins for human roles in certain contexts (e.g., virtual assistants, customer simulators). 

Insight Extraction Applying AI or ML to identify and explain meaningful patterns in large datasets, helping brands uncover new opportunities or consumer insights. 

AI-Driven Decision Making The use of algorithms or AI models to assist in strategic or operational decisions by predicting outcomes or assessing risks. 

Human Augmentation Enhancing or extending human capabilities with AI, such as digital twins, augmented analytics, or brain–computer interfaces. 

Data Set A collection of information that AI uses to learn patterns and make predictions. 

Bias When AI produces skewed or unfair outputs due to the data it learned from. 

Agent (Agentic AI) An AI system that can act autonomously or semi-autonomously to achieve goals, rather than just responding to a single prompt. Can plan, act and iterate, often performing multi-step tasks such as research, content generation or campaign optimisation. Think of it as a “virtual team member” rather than a single-shot content generator. 

Fine-Tuning Adjusting a pre-trained AI model on specific data so it performs better for a particular task, like generating brand-specific copy. 

Generative Prompt Engineering Crafting effective inputs to maximise the quality and relevance of generative AI outputs. This often requires iterative testing. 

Style Transfer An AI technique that applies the style of one image or creative asset to another, for example: making a photo match a brand’s visual identity. 

Hyper-Personalisation Using AI to create individualised ads, emails or product recommendations at scale. 

Content Moderation AI Automated systems that detect inappropriate or unsafe content before it reaches audiences. 

Ad Creative Optimisation Using AI to test and improve multiple variations of ad creatives automatically. 

Predictive Analytics AI-driven forecasting of consumer behavior, campaign performance or media outcomes. 

Image-to-Text / Text-to-Image AI tools that convert images into descriptive text or create images from written prompts. 

Synthetic Media AI-generated media assets that mimic reality, from voices to actors to video backgrounds. 

Synthetic Audiences AI-generated models of audience groups that simulate real consumer behaviours, preferences and responses. Used to test creative, messaging or media strategies without relying solely on real-world data. Synthetic audiences help brands predict outcomes, experiment safely and optimise campaigns before launch. 

Explainable AI (XAI) AI systems designed to provide transparent reasoning for their outputs. 

Large Language Model (LLM) Advanced AI trained on massive text datasets to generate coherent, human-like language. Examples: GPT-4, Claude, Bard. 

Transformer Architecture The underlying neural network design powering modern LLMs and generative models.  

Reinforcement Learning with Human Feedback (RLHF) Training AI models using feedback from humans to align outputs with preferred behavior, improving safety and relevance. 

Multimodal AI AI capable of understanding and generating content across multiple modalities, for example text, image, audio and video simultaneously. 

Prompt Injection A security risk where malicious inputs trick AI into executing unwanted instructions. 

Zero-Shot / Few-Shot Learning When AI performs tasks it has not explicitly been trained on, with little or no examples provided.  

Neural Rendering An AI technique generating high-quality images or video by simulating the physics of light, materials and movement.  

Generative Adversarial Networks (GANs) Two-part AI systems that compete to create realistic outputs; foundational for synthetic media and image generation. 

Content Attribution AI AI that can automatically tag, trace and protect brand content across platforms for IP and copyright management. 

AI Ethics / Responsible AI Frameworks and practices to ensure AI is used fairly, safely and transparently, mitigating bias, discrimination and reputational risk. 

Adaptive AI Strategy A business approach focused on embedding adaptability across AI applications - combining automation for short-term efficiency with adaptive systems for long-term resilience and competitive advantage. 

Responsible Adaptation The principle of designing adaptive AI systems that are ethically grounded, explainable and secure from the outset to build consumer and brand trust. 

Brand Safety AI Platforms that automatically flag risky content contexts for ads (e.g., ad placed near misinformation or offensive content). 

Digital Twin Platforms AI-driven simulations of real-world systems, customer journeys,or marketing ecosystems. Adaptive AI allows brands to test scenarios and predict outcomes before implementation. 

Adaptive Analytics Suites Tools that continuously learn from campaign data, audience behavior, and operational metrics to optimise marketing strategy and decision-making over time.  

Human-AI Collaboration Platforms Systems designed to integrate domain experts and AI agents, enabling iterative co-creation, adaptive content and real-time decision support. 

Augmented Marketing Platforms Tools that enhance human decision-making with AI insights, predictive models and adaptive suggestions for campaign optimisation, content personalisation, and resource allocation. 

 

Brand tool / Platform terms

ChatGPT / Bard / Claude AI tools for generating text-based content, idea brainstorming and conversational applications. 

DALL·E / MidJourney / Stable Diffusion Generative AI tools for creating images from textual prompts. 

Runway / Synthesia AI platforms for video creation, editing and synthetic media, often used to scale production or create hyper-personalised campaigns. 

Copy.ai / Jasper AI tools focused on marketing copywriting and content ideation. 

Surfer AI / MarketMuse Tools that combine AI with SEO/marketing insights for optimized content strategy. 

Canva AI / Adobe Firefly AI-assisted creative tools for image generation, layout and visual experimentation. 

Lumen5 / Pictory Video creation tools that turn scripts or blog posts into branded video content automatically.