Abstract
Managing personal finances remains a significant challenge in developing economies, where access to financial advisory services is limited and financial literacy is often low. In response to this, we propose «ECO InvestMind AI», an intelligent mobile application designed to democratise wealth management through the integration of advanced machine learning (ML) and natural language processing (NLP) techniques. Our system leverages a dual-stream hybrid architecture combining a fine-tuned BERT model for sentiment analysis of financial news and an MLP-based time series model for stock price forecasting. The application provides personalised stock recommendations by analysing market trends, Environmental, Social and Governance (ESG) scores and sentiment signals, offering users real-time, behaviourally adaptive financial guidance. To ensure scalability and maintainability, the app is built using Clean Architecture principles and Kotlin Multiplatform, enabling seamless cross-platform performance. A relational database structure supports the integration of diverse datasets, including historical stock prices, ESG metrics and sentiment-labelled news. Evaluation results demonstrate high predictive accuracy across multiple Kazakhstani companies and indices. This paper illustrates how AI-powered tools like «ECO InvestMind AI» can bridge the gap between traditional investment tools and modern financial advisory systems, empowering users with actionable insights and promoting inclusive participation in financial markets. The proposed hybrid algorithm achieved a total system accuracy of 96.87%, confirming the effectiveness of combining sentiment analysis and time series forecasting for intelligent investment decision-making.