Chris Ackerson, SVP of Product at market intelligence platform AlphaSense, noted that general large language models (LLMs) exhibit an error rate of 5% to 9% when using leading market data in the financial sector, which is unacceptable for investment professionals. He emphasized that vertical AI, through extreme control over data, can provide evaluation results three times better at a lower cost. AlphaSense has built a proprietary database containing over 300,000 expert interview transcripts and developed an AI interview system whose interview quality has met or even surpassed that of the best analysts. Ackerson revealed that proprietary search models can reduce financial research costs by up to 40 times. He also stated that AI has not eliminated financial practitioners but has increased their average coverage by 15%, allowing them to focus on higher-value activities.