AI.FORECAST

AI.FORECAST refers to a category of products where AI predicts future numbers—like sales, visitor counts, or inventory levels—by looking at several related pieces of information at the same time. It's similar to a weather forecaster who doesn't just look at temperature alone, but combines humidity, wind, and past weather records to judge whether it will rain tomorrow.

In the past, many forecasting methods took a single number, like sales, and simply extended the pattern seen in past data into the future. But in a real store, multiple signals—promotion schedules, sales of related products, weather forecasts—affect the outcome at the same time. AI.FORECAST-type products take in all these signals together, calculate how they influence one another, and instead of giving one fixed number, they answer in the form of 'within this range, this outcome is most likely.'

So when looking at these products, it's worth checking whether they simply give you a single number, or whether they consider multiple data sources at once and also show the uncertainty involved.

In news articles, you might see something like 'Google Research released TimesFM-3, a foundation model for multivariate time-series forecasting.' A common misunderstanding is that AI.FORECAST is not the name of a single product from one company—it's a general term for the whole category of models and services that predict the future by feeding in multiple data streams at once.