AI agent decision-making process

The AI agent handles:

• Rule translation (e.g., “Buy if ETH < $2700” → on-chain condition)

• Strategy optimization (multi-protocol routing)

• Backtest simulation using historical subgraph data

Example AI-compiled strategy (simplified):

{
  "protocol": "uniswap",
  "action": "swap",
  "tokenIn": "USDC",
  "tokenOut": "ETH",
  "trigger": {
    "type": "priceBelow",
    "value": 2700
  }
}