FractionAI has commercialized an index solution designed to comprehensively support AI agent-based asset management.
The company said the platform goes beyond conventional automated trading by integrating the full workflow—from natural-language strategy design to execution, validation and learning. Rather than applying only pre-defined algorithms, the core function is converting a user’s trading idea written in everyday language into a strategy that can operate in real markets.
Using advanced natural-language processing, the system analyzes narrative instructions and turns them into quantified investment rules. Elements such as entry timing, capital allocation and loss tolerance are not set in isolation but combined into a single logical structure so the overall strategy remains consistent. FractionAI said the approach is intended to lower the barrier for users who previously needed to code or enter complex formulas, while maintaining precision in strategy design.
Before commercialization, the company ran simulated tests to check the conversion algorithm and validation framework. Participants were able to watch in real time as natural-language ideas were implemented as trading rules, helping verify accuracy and consistency. The process also served to refine user experience and confidence in the algorithm, the company said.
Strategies are not deployed immediately. They first undergo simulated trading, or backtesting, using historical time-series data accumulated on the platform. In this stage, the system analyzes not only profitability but also loss volatility and vulnerabilities under specific market conditions. Users can revise and refine a strategy based on the results, then complete a final approval step before activating an automated trading agent—linking design, verification, improvement and execution in a single flow.
After deployment, the system continues to collect data. Trading outcomes are recorded in detail, including profits and losses and response patterns across market phases, and are used as learning data for future improvements. FractionAI said this feedback structure is intended to support an adaptive system that can strengthen strategies over time, rather than simply executing trades automatically.
The solution is designed for the cryptocurrency futures market. With high volatility and 24-hour trading, rapid execution and risk management are critical, making automated decision systems particularly useful. FractionAI said it built the platform to combine immediate strategy updates with continuous data-driven learning suited to that environment.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.