Good day,
On making objects out of records, artificial intelligence may be used to classify the record as the name given. The object may be turned into a numpy array and stored with the label. A linear regression may be run with the objects under the same label as predictors and new entries as dependents. Iteration of the regression may create an accurate model enough to predict new records. On entry, the record may be turned into an array and made dependent for probability. A flag may be given if not greater than or equal to the model inference. If so be it, autocorrection may be done on record entry by iterating through the array and suggesting the record entry. This may be artificial intelligence in the realm as a model.
If done, prediction may be done on the objects for analytical purposes. That is, if thought goes so far.