LG, London Stock Exchange unveil AI equity forecast tool

LG and the London Stock Exchange Group (LSEG) have launched an AI-powered equity forecast score service that uses machine learning and financial data to predict market trends and assess risk.

The service, developed by LG AI Research and LSEG, combines LG’s AI tools with LSEG’s financial datasets to analyze thousands of US-listed stocks daily.

It provides a numerical score predicting expected returns over a four-week period, and a text commentary explaining the forecast.

The system uses deep learning to process both structured market data and unstructured financial documents, covering companies of various sizes, including nano-cap stocks.

Institutional clients are currently testing the service, which updates scores daily and releases weekly commentary.

LG plans to expand the tool to forecast equities in other regions and add features for portfolio structuring and commodities.

🔗 Source: The Korea Times

🧠 Food for thought

Implications, context, and why it matters.

Missing performance metrics leave institutional adoption uncertain

LG and the London Stock Exchange Group (LSEG) launched an AI equity forecast score that predicts four-week stock returns 1. Institutional investors need validation metrics before they try a signal. The product does not publish out-of-sample returns (performance on unseen data) or Sharpe ratios (risk-adjusted returns). It also omits hit rates (percentage of correct signals) or turnover costs (trading frictions from frequent rebalancing).
For nano-cap stocks (very small companies by market capitalization) transaction costs can erode predicted returns. There is no sign the forecasts account for bid-ask spreads (the gap between buying and selling prices) or market impact (price moves caused by executing trades).
LG AI Research’s EXAONE-BI (its multi-agent business-intelligence model) uses four specialized AI agents that work together 1. No independent test data covers accuracy across market conditions. That leaves clients unsure whether the service delivers alpha generation (returns above a benchmark) or pattern recognition that may fade.

Application Programming Interface (API) access could open distribution paths for fintech builders

LG and LSEG have not shared how the score will be distributed or if an API will exist 1. LSEG offers the Refinitiv Data Platform APIs 2. LSEG operates in 65 countries 1, which could open distribution if the score lands in those APIs.
If LSEG adds it, developers could plug the 1-100 score with LG’s commentary into risk dashboards, portfolio tools, or screening apps 1. They could reuse API libraries, documentation, or authentication methods 3 4.
Developers can layer the AI score with datasets in LSEG’s API catalog 5. Options include ownership data, analyst estimates, or company fundamentals 2. That mix can power decision-support for wealth managers and asset allocators who want contextualized signals over raw predictions.