Reflections on the Future of Finance: From Quants to AI Agents
My take on how AI and democratization will change finance for investors and learners alike.
Introduction: Gratitude and Momentum
While I am still recovering from the long weekend fatigue, I want to thank everyone who reached out and shared feedback on my last blog post. I am grateful for your encouragement and excited to keep learning about the companies and products working to modernize investment and quant research. I’ve made meaningful connections and I am looking forward to experimenting with new products that could help streamline my workflow.
Sharing My WorldQuant University Interview
Besides teaching as an adjunct instructor at WorldQuant University, I was honored to sit down with Greg Ciresi for an interview about my career journey, from traditional finance to fintech to education. I also shared my excitement about experimenting with various AI tools for investment research.
We discussed the new skills modern quants should have, including being AI-native, interdisciplinary, and proactive in driving initiatives within organizations. I particularly loved what Andrew Ng said at this year’s Snowflake Dev Day: Move fast and be responsible. This is exactly what modern quants should do. We move fast, use AI to rapidly prototype ideas, and at the same time understand that disruption must also be responsible. We are working with investment strategies and helping investors, both large and small, plan for a sustainable financial future.
The Three Pillars of Democratizing Finance
I also shared my views on three pillars of democratization in finance:
Everyday investors gaining access to personalized, high-quality financial advice
Rising/Emerging managers and solo advisors using open-source and AI tools to compete
Global learners (including WQU students) gaining access to elite education
We already see many companies working to make investment more accessible. I am particularly passionate about how AI, open-source solutions, and open data can help smaller funds compete with larger shops. Lastly, I believe financial education should be made more accessible. WorldQuant University is one of the best examples of this, and I am excited to be part of it. Even within a few weeks, I can feel the passion and energy of the students.
The Future of Hyper-Personalized Financial Advice
As I have mentioned before, I am fascinated by the idea of a hyper-personalized all-in-one financial solution. In the future, we will no longer need to work with multiple professionals such as an accountant, insurance agent, financial advisor, or mortgage broker. Instead, a single intelligent system could generate scenario analysis to help consumers navigate different milestones in life, predict personal changes, and adjust financial behavior accordingly.
Key Takeaways from “The Dawn of AiFi” Webinar
Speaking of personalization, I recently attended the webinar The Dawn of AiFi hosted by Ryan Favley at Restiv Ventures. AiFi stands for AI in Financial Services and is described as a new financial system designed for AI agents.
The session highlighted how agents could soon handle day-to-day financial activities such as approving transactions, managing micro-sized payments, and protecting consumers with advanced security measures. Ryan also shared a few interesting startups, two of which stood out to me: Hiro and Reload.
Startups to Watch: Hiro and Reload
Hiro positions itself as a “personalized CFO” that provides intelligence for financial decisions. Features include spending analysis and cash optimization. I am excited to see a company working on personalized financial intelligence and will be following Hiro closely. As mentioned earlier, I believe the future lies in one-stop solutions for all financial needs.
Reload describes itself as the first payroll system for AI agents. This caught my attention because while we usually think about large organizations building in-house agents, there are now marketplaces where people can browse and buy agents. I have been exploring these marketplaces, hoping to find financial-service-focused agents. While I haven’t seen many yet, I plan to keep exploring. My next step will be to experiment with both in-house agents and generic marketplace agents, comparing the pros and cons, particularly in terms of implementation difficulty, operations, and cost.
Exploring Different Approaches to AI-Assisted Coding
I also came across an insightful blog post by Sabestien Castiel in his blog Between the Prompts, where he categorized three approaches to AI-assisted coding:
Vibe Coding – asking an LLM to build an application entirely via chat, often used by non-technical professionals to quickly prototype ideas.
Copilot Approach – using an LLM like a junior employee, offering instructions to build parts of an application under supervision.
Tool-Based Approach – using LLMs solely as coding tools. For example, Cursor’s tab feature helps developers write code more efficiently.
Personally, I fall somewhere between the second and third approaches. When I use LLMs as copilots, I treat them like interns, requiring extremely detailed instructions to avoid bugs and mistakes. For simple tasks or utility features, I lean toward the third approach.
I am always looking for creative ways to use LLMs, so I am glad to see more people analyzing and organizing these schools of thought.
Closing Thoughts
It is an exciting time to be both a quant and an educator. I will continue exploring how AI, open-source tools, and agents can democratize finance while making education and research more accessible to all.
On AI coding approach #2, I found myself increasingly spend more time discussing bigger scope plans and detailed tests at the beginning, and spend less time giving instructions on mid-course implementation choices. I am still learning how to design and articulate testing and evaluation better.