What is machine learning?

Machine learning (ML) is a way of building software that learns from examples instead of following rules written by a person. It studies many past cases, finds the patterns that link inputs to outcomes, and uses them to make predictions about new cases. Spam filters, fraud checks, sales forecasts and credit scores use machine learning. So do the AI models behind ChatGPT and Claude.

Why it matters for a business

The word now covers 2 very different things. Classical machine learning is a model trained on a company’s own structured data, such as rows of past orders or closed deals, to predict a number or assign a category. It answers in milliseconds, costs almost nothing per prediction, and can show which inputs moved the result. Large language models are machine learning too, trained on public text to read and write language. When an engineer says “use ML for this, not an LLM”, they usually mean the first kind fits the problem better.

Example

Let’s say a distributor wants to know which customers are likely to pay late again. A classical model trained on 3 years of the company’s payment history scores every account in milliseconds. A language model has never seen that history, so any ranking it gives is a guess based on businesses in general.

Go deeper

AI vs. machine learning, and where LLMs fit

Related terms

• Deep learning
• Neural network
• Large language model (LLM)
• Generative AI
• Training data

All terms, A to Z