Machine learning
Machine learning is how computers learn to do tasks from examples, without rules written by a person. Today's AI platforms are built with it.
Machine learning is a way of building software that learns from examples, with no rules written by a person. Show it thousands of emails marked spam or not spam, and it works out for itself what spam looks like. The AI models behind ChatGPT, Gemini and the rest are built this way, trained on huge amounts of text.
For a marketer, the useful part is what follows from that. Nobody wrote a rule that says "name this brand." The model learned patterns, and the patterns decide.
How it works
You give the system examples and a goal. It makes a guess, checks how far off it was, adjusts itself a little, and repeats, millions of times. What's left at the end is a model: a large set of numbers that captures the patterns in the examples.
A language model's goal is to predict the next word in a piece of text. Get good enough at that across enough text, and it can answer questions, write summaries and compare products.
What it means for your brand
Because the model learns patterns, it repeats what it saw often. A brand that appears in lots of reviews, articles and forum threads becomes part of the pattern. A brand that rarely appears doesn't.
It also means you can't ask the platform why it named your competitor and get a real reason. There's no rule to point at. You can only look at what it read and what it said.
For example, if most of the articles about coffee subscriptions mention Kestrel Roasters alongside "fresh," a model may learn to connect the two. Ask it for "the freshest coffee subscription" and Kestrel comes to mind first, even if Fernhill Coffee roasts more recently.
Learning and searching
Training happens before you ever ask a question, and it's slow to change. That's why most AI platforms now also search the web while answering, so what they read today can add to what they learned. That second part is where you can make a difference quickly.