Supervised learning: Supervised learning trains a machine learning model on examples that already carry the correct answer, so it can predict the answer for new cases. It is behind spam filters, credit scoring, demand forecasts, and most business AI that classifies or predicts.
The name comes from the teacher: a labeled dataset that tells the model what the right output is for each input. Show it a million emails marked spam or not spam, and it learns which patterns of words, senders, and links predict spam. Show it five years of sales by store, week, and weather, and it learns to forecast next month. The model adjusts itself to shrink the gap between its guesses and the labels.
Two flavors cover most business uses. Classification picks a category: fraud or legitimate, churn risk or not, which department a ticket belongs to. Regression predicts a number: how many units, what price, how many days until payment. Either way, the model only knows what its examples taught it. Feed it data from a different era, region, or customer type and its accuracy can fall off quietly.
Large language models complicate the picture slightly. Their initial training is self-supervised: the "label" is simply the next word in existing text, so no human tagging is needed. But the steps that make them useful assistants, including fine-tuning on example conversations and learning from human ratings, are supervised in spirit. When a vendor says a model was "trained on your data", supervised fine-tuning is usually what they mean.
Example at work
A property management company trains a model on three years of maintenance requests labeled by urgency. Now every new request gets an urgency score before a human sees it, and emergencies jump the queue. Twice a year the team reviews the requests the model scored wrong and adds them to the training set.
Why it matters
Most AI that runs a business quietly, from lead scoring to fraud alerts, is supervised learning. Understanding that it learns from labeled history explains both its strengths and its blind spots: it is only as good, and as current, as the examples it was given.
Related terms
- Unsupervised learningUnsupervised learning is machine learning on data with no correct answers attached. The model finds structure on its own, grouping similar customers, spotting unusual transactions, or clustering tickets by topic; a human then decides what the groups mean.
- Machine learning (ML)Machine learning is the branch of AI in which software learns patterns from examples instead of following rules a programmer wrote by hand. Nearly every modern AI tool, including chat assistants, is built on it.
- Data labelingData labeling is the work of attaching correct answers to raw data, such as tagging emails as spam, drawing boxes around pallets in photos, or rating chatbot replies, so a machine learning model can learn from those examples.
- Training dataTraining data is the collection of examples an AI model learns from, such as web pages, books, code, images, or labeled records. What is in that data, and what is missing, shapes everything the model knows, how it writes, and which mistakes it makes.
- Fine-tuningFine-tuning is extra training that adjusts an existing AI model's weights using a smaller, targeted set of examples so it performs better on a specific task, style, or domain. It changes the model itself, unlike prompting, which only changes the instructions.