In my conversations across retail investing forums, a clear theme has emerged: while excitement around AI is high, the dense jargon surrounding "Agentic AI" and LLMs is leaving many investors on the sidelines.
My goal today was to strip away the complexity. I wanted to give investors a jargon-free toolkit to interact with Large Language Models, understand their capabilities across different price points, and optimize their research workflows. While we focused today on Claude CoWork and Anthropic models in particular, the core concepts are applicable across AI platforms.
We broke this down across three sessions:
The Foundations: Demystifying the nuances, capabilities, and common pitfalls of LLMs.
The Progression: Moving step-by-step from basic prompting to structuring Standard Operating Procedures (in the form of Skills), experimenting with Agents (for specific goals), and orchestrating everything through a central research assistant (meet ARYA!) To keep things practical, I used detailed companion notebooks for both Mac and Windows, with outputs captured there for all prompts. You can find the detailed companions here
The Art of the Possible: In the final session, we looked at the endgame: building a sophisticated army of agents capable of doing the heavy lifting for multi-dimensional, deep equity research.
We had good participation, and the feedback has been incredibly encouraging. I am uploading the slide deck here for the community. Link to the presentation The recording for the session will be made available shortly. I will share the link here. I would love your feedback and suggestions—please share your thoughts on how we can continue to develop and refine these agentic systems together. Happy Learning and (Agentic) research-led Investing!
