Why is AI getting better at finance, yet top analysts aren’t becoming any less valuable? In the past, institutions were willing to pay a high cost for research reports, with much of that expense going toward reviewing financial statements, compiling meeting notes, building models, and preparing pitch materials. Now, with OpenAI’s ChatGPT for finance, which can connect to diverse professional data sources, these tasks can be completed in a fraction of the time. As a result, the first thing to lose value may be “presenting information beautifully.” But finance has a unique quality: every report ultimately ties to real money. A model can calculate ten scenarios but won’t bear the burden of the eleventh unexpected event; it can deliver a complete conclusion but doesn’t need to explain to clients afterward which assumption failed when losses occur. After years of trading, I’ve come to believe that in finance, the most valuable thing has never been the answer—it’s someone willing to take responsibility for their judgment. In the future, information will become cheaper, and professional reports will grow increasingly similar. The real difference between people will lie in their ability to ask the right questions, identify overlooked risks, and still dare to reject an appealing conclusion—even when all the evidence looks perfect. AI can make every report look professional. But whether it’s worth betting real money on it? That still requires a human to sign their name.