MSCI has rolled out SignalLab, a new platform that combines governed, research-based signals from several investment domains with the tools investment teams need to test and deploy them at scale.

The platform responds to a growing problem for investment teams. Signals now come from many vendors, each with its own quality standards, methodologies, investment universes and security identifiers, which makes them difficult to compare and slows down onboarding. Firms have generally had two options: develop signals in-house, which tends to be slow and costly, or piece them together from multiple suppliers. SignalLab instead offers a curated library of standardised, validated and documented signals in a single environment supported by MSCI’s data and infrastructure. Clients can also add their own proprietary data to the same environment.

SignalLab includes more than 600 research-grade signals with global coverage, refreshed daily across four areas. The first, risk premia, covers signals rooted in economic and financial literature. An AI agent reviews the research and converts it into tested, scored signals, and an MSCI researcher then checks them. The second, micro-industries, is a granular, AI-built company classification. A proprietary language model performs the initial clustering, large language models sharpen the definitions and allocate stocks, and MSCI’s model validation team reviews the results. The remaining two areas are MSCI’s existing FactorLab and Crowding signals.

MSCI plans to expand the platform with further signals and data modules that support signal customisation and connect data with its index and risk model production environments. Two modules are in the pipeline.

Source link