FinSights

This AI agent for investing delivers answers backed by New Constructs’ proprietary dataset, built from more than 20 years of analyzing millions of financial filings. A patented system for reverse-engineering accounting distortions gives an accurate view of a company’s true profitability and valuation, and the same data powers New Constructs’ Bloomberg indices.

Investors can ask for the best stocks in a sector or the stocks most likely to miss or beat earnings next quarter. A natural language screener finds stocks, ETFs, and mutual funds that meet specific criteria, such as undervalued technology stocks with high return on invested capital (ROIC) and positive free cash flow. Users can get an unbiased view of any company’s financial health, valuation, and risk profile, and compare large numbers of companies by asking natural investor questions. It runs on BigQuery and Gemini models for fast analysis of large datasets, and its data already drives trading strategies for institutional clients.