Purpose - This paper explores how financial inclusion depends on demographic factors such as, age, gender, education, occupation and income, in the Sambalpur and Bargarh districts of Odisha, India, and uses the latest machine learning to study the subject.

Design/Methodology/approach –The multistage sampling was conducted on a population of 813 respondents, of which 70 percent were used as a training sample, and 30 percent as a testing sample. Different models were used to analyze the data, among them are “decision trees, artificial neural networks (ANN), regression trees, regression model and deep neural networks (DNN)”. “Root Mean squared error (RMSE) and Mean Absolute errors (MAE)” were used as measurements of performance.

Findings - Second ANN model had high performance in forecasting financial inclusion as the RMSE and MAE of the training and testing data are lower. Moreover, the 1st, 3rd, and 6th DNN models were very accurate in boosting financial inclusion prediction. The analysis of the decision tree showed the high income is a very strong predictor of financial inclusion, but other attributes like age, occupation, and gender complicate the situation with the majority of the population. The regression tree also found the income as the strongest variable, whereas the relationships were more nuanced, including the age, occupation, and education variables, on the lower-income groups.

Originality/Value - The reading has added worth to current works on applying AI-based approaches to forecasting financial inclusion as it reveals its possible usage in other geographical markets..