Previously, the focus was on the strength of the model; now, people are more concerned with what AI can accomplish in real-world business. Trading is one of the most direct testing scenarios because every step involves real money. On November 10th, @NeoSoulAI released NeoConnect. This is a trading engine for platforms, enterprises, and AI agents, along with FDE enterprise AI services. The first half of AI trading involves identifying opportunities and determining order motivation. Models for this part are becoming increasingly sophisticated, and AI making independent decisions is no longer difficult. The challenge lies in the second half: how to place orders, control total position size, manage risk, and conduct post-trade analysis. NeoConnect sells this part. Traditional brokerages are also moving in this direction. IBKR has upgraded its previous model of AI proposing and human approval to authorizing AI to automatically place orders in separately funded sections, covering over 75,000 stocks, options, and ETFs. Brokerages are willing to entrust execution to AI, provided they first use a separate fund partition to contain the risk. This indicates that execution under controllable conditions is the real barrier to entry in this field. From a commercial perspective, the closer to the order placement step, the more justifiable it is to charge for off-the-shelf solutions, because the platform's other option is to maintain its own team. NeoSoul's approach has two layers. The first layer is the standard engine, which the platform can use immediately upon connection, allowing for rapid deployment. The second layer is FDE, which is the frontline deployment engineer model commonly used by Palantir. Once a client integrates their internal system, it's difficult to switch; the cost is that each client requires an engineer, making delivery more cumbersome and expansion slower. On the product side, NeoTrade Web launched on November 3rd, and NeoConnect was released on November 10th. Trading volume reflects token popularity, and product launches reflect team delivery. Neither of these lines equates to commercial revenue. What connects these two lines is the integration partners. The number of partners integrated, their types, and the actual transaction volume will tell the market whether anyone is willing to pay for this business.