Playgent

Playgent builds reinforcement learning environments for finance, used to post train and evaluate AI agents on realistic financial work. A reinforcement learning environment is a simulated world in which an agent takes actions, observes the resulting state and receives rewards, and Playgent applies that idea to tasks in investment banking, compliance and operations. The company works with a network of finance experts who curate difficult tasks, design verification rubrics and build the environments. Example tasks on the site include a leveraged buyout returns analysis for a large takeover, normalizing earnings to produce adjusted EBITDA for a real estate investment trust take private, and a synergies and integration analysis for a regional bank merger. Each task is benchmarked, and the vendor sells the resulting tasks, verifiers and supporting infrastructure as production ready environments for post training. It also publishes research on improving agents at hard investment banking tasks. The offering is aimed at AI labs and teams building finance agents who need high quality training and evaluation data rather than a finished end user application. Getting started means contacting the company to discuss the use case and access the environments. Pricing is not published. The site has no pricing page, and the vendor does not list plan prices or data rates.