What is machine learning?

Definition

Machine learning is the part of AI where software learns from examples instead of being programmed step by step. You show it lots of past data (emails labeled spam or not, accounts that did or didn't churn) and it learns patterns it can apply to new cases.

Generative AI is built on machine learning, but when people say "machine learning" in a product context they usually mean predictive features: scores, rankings, alerts and classifications.

Why it matters for advisors

Many tools advisors already use run machine learning quietly in the background: surveillance alerts in archiving tools, risk flags in compliance review, lead scores in prospecting databases and categorization in document tools.

A prediction is only as good as the data behind it. If a model learned from a different kind of client base, its scores can be off for yours. FINRA's 2026 oversight report lists bias from limited or outdated training data as a risk firms should consider.

Example

A prospecting tool ranks leads by how likely they are to need an advisor. Before relying on it, the firm asks the vendor what data the model learned from and checks a sample of top-ranked leads against its own idea of an ideal client.

Watch out for

A score looks precise but isn't a fact. Treat machine learning output as a signal for a person to review, especially where it affects clients.