What is generative AI?

Explained

Generative AI is a family of models that produce new content in response to an instruction: text, images, code, audio or video. Instead of assigning a label to an input, the model learns the patterns of very large datasets and generates an output that follows them. It powers chat assistants such as ChatGPT, image generators, code assistants and a growing number of enterprise applications. Its strength is flexibility; its weakness is that outputs can be fluent and still wrong, so it needs grounding, evaluation and human review where the stakes are high.

How it works

Most generative AI today is built on large neural networks trained on huge amounts of data. A language model learns to predict the next token of text; repeated many times, that produces paragraphs, answers or code. Image and audio models use related techniques to turn a description into pixels or sound.

The same model can do many tasks because the task is described in the prompt. In enterprise use, the model is usually combined with company data (for example through retrieval), tools it can call, and rules about what it may and may not do.

What businesses use it for

• Drafting and summarising documents, emails and reports.
• Answering questions over internal knowledge, policies and manuals.
• Customer service assistants and contact centre support.
• Writing, reviewing and explaining code.
• Extracting structured data from contracts, invoices or forms.
• Creating images, video or voice for marketing and training.

Generative AI vs traditional AI

Traditional machine learning is mostly discriminative: it classifies, scores or forecasts, such as flagging fraud or predicting demand. Generative AI produces something new: a reply, a draft, an image.

The two often work together. A predictive model can decide which customers need attention, and a generative model can write the message. Generative AI does not replace analytics; it adds a language and content layer on top.

Limits to plan for

• Hallucinations: answers that sound right but are not supported by facts.
• Data exposure: prompts and outputs can contain confidential or personal data.
• Cost and latency, which grow with model size and usage.
• Intellectual property and regulatory questions, including obligations under the EU AI Act.
• The need for evaluation and monitoring once the system is in production.

Frequently asked questions

Further reading

• Generative AI Strategy: Implications and Opportunities for Businesses
• Generative AI: From Hype to Real-World Value Creation

Related terms

• Evaluation The process of measuring how well an AI model or application performs: accuracy, relevance, safety, bias, cost and other criteria.
• Grounding Grounding ties a model's answers to verifiable sources, such as documents, databases or search results, instead of relying only on what it memorised.
• Agent Agenda An AI agent's live plan: the ordered next steps, open subtasks, blocked items and checkpoints it works through towards a goal.
• Agent Planner The component or prompt pattern in an AI agent that breaks a goal into a sequence of steps before any tool is run.
• Agent Skill A packaged capability for an AI agent: the instructions, allowed tools, examples and input and output formats it needs to perform one coherent task.
• Agentic AI AI systems that pursue a goal with some autonomy: they plan steps, use tools and act on systems, rather than only answering a prompt.