Skip to content

AIM Investment Process

core

AIM takes four key steps that seamlessly combine realtime data collection, LLM-based analytics, and automated trading decisions to continually enhance its feedback loop.

Data Collection

  • Crypto News: Realtime updates on major economic events, industry trends, and project announcements
  • Market Analysis: Exchange price fluctuations, trading volume, order book data, and volatility indicators
  • Previous Decisions: Historical records of past trades and their outcomes for ongoing learning
  • Fear and Greed Index: A measure of market sentiment to gauge crowd psychology
  • Current Investment State: Existing portfolio positions, transaction history, and internal analytics

All these inputs create a comprehensive data environment for subsequent LLM-based evaluations.

Request Investment Options

  • GPT O1, Claude 3.5, and Deepseek R1 each analyze the aggregated data independently.
  • Recommendations can include Buy, Sell, Hold, or other strategic insights.
  • Different models bring diverse perspectives, reducing the likelihood of tunnel vision in making decisions.

Final Decision

  • The individual outputs from the LLM models are consolidated into a single decision.
  • By leveraging multiple sources, AIM aims to arrive at a balanced, well-rounded judgment.
  • This final outcome directly informs the next stage of trading execution.

Trade Execution

  • Based on the unified recommendation, AIM executes Buy, Sell, or Hold orders in real time.
  • Trade outcomes (profits/losses, updated positions) are stored in a database for future reference.
  • These results feed back into the Data Collection phase, closing the loop and informing subsequent decisions.

Continuous Feedback and Growth

By repeating these steps, AIM ensures that previous outcomes constantly refine its models and strategies. As new data accumulates, the system adapts to evolving market conditions, aiming for higher performance and sustained value creation over time.