This is part 2 of a series on leveraging LLMs in your fundamental analysis workflow. In Primer 1, we collected ideas from Sellside/Buyside outlooks and used ChatGPT to synthesize an alpha capture outlook for global stocks. This series is intended for fundamental and discretionary research analysts who are seeking to incorporate quantitative methods into their research process. As a practitioner of over 20 years who's also made the journey, I'm sharing my research process step-by-step to demonstrate the power of these tools to build and test an investment thesis. Perhaps more importantly, we will demonstrate how to build a process around sanity checking to ensure that the model’s output reflects your expert opinion. As we conclude this piece, we reiterate that Computers probably shouldn’t “assume” anything about the future direction of Central Bank policies!