The Future of Finance in the Age of AI

Key Concepts

• Generational Shifts in Consumer Behavior: Millennials (internet in households), Gen Z (iPhone and social media), Gen Alpha (AI generation).
• AI as a New Computing Paradigm: Distinct from previous technologies, changing human-computer interaction.
• Personal AI Application: Using AI to solve personal problems as a learning pathway before business application.
• Data-Driven AI: The critical role of data in training and utilizing AI, whether personal health data or financial data.
• AI Agents: The next evolution of AI, acting as co-pilots to accomplish tasks using various tools.
• AI Value Chain: Infrastructure (GPUs, power), Large Language Models (LLMs), Data, and Applications (text, video, voice, visualization).
• End of the Knowledge Economy: Shift from memorization and regurgitation to creativity, problem-solving, and critical thinking.
• Future-Proofing: The necessity of understanding and building with AI to remain relevant in the evolving job market.
• AI Agents in Finance: Expected to become routine in operations within three years, aiding in forecasting and strategic decision-making.
• Talking to Data: The ability to query and understand financial data through natural language interfaces.
• Business as a Math Equation: Leveraging AI to quantify business processes and drive strategic change.

The Evolution of the Consumer and the Rise of AI

The speaker, a Gen Xer, has spent their career helping brands understand evolving consumer generations. Initially, the focus was on millennials, the first generation to grow up with the internet in their homes. This was followed by Gen Z, characterized by their deep integration with smartphones and social media, often described as the "iPhone generation." Now, Generation Alpha (currently aged 0-15) is emerging as the "AI generation," who will never know a world without AI and will interact with technology as naturally as they do with people. This generational shift highlights a fundamental change in how humans engage with technology.

Personal Journey into AI: Solving a Personal Problem

The speaker's venture-funded software company, Suzie, initially struggled to integrate AI. The realization that AI represents a new realm of human-computer interaction led the speaker to take a personal approach. As a 50-year-old with young children, the speaker's personal goal is to "stay alive as long as possible." This led to the creation of a personal health bot trained on 25 years of personal health data (MRIs, X-rays, blood tests, doctor's notes). The bot, instructed to act as a leading doctor from Johns Hopkins, provided "incredible insight" by identifying potential issues from past blood tests and offering a dossier for medical appointments. This experience underscored the power of AI to provide custom, data-driven insights, contrasting with general health information sites like WebMD.

Overcoming Barriers to AI Adoption

For many in larger companies, privacy and data security are significant encumbrances to adopting AI tools and uploading customer data. The speaker's advice to overcome this is to focus on a personal problem first. Examples include managing personal finances, trusts, or even helping a parent with simple tech issues like password retrieval. The process involves:

1. Identifying a problem to solve.
2. Aggregating the necessary data.
3. Building a solution.

The speaker emphasizes that this process is "incredibly easy" and can be guided by AI itself, for instance, by asking ChatGPT for step-by-step instructions. The key learning is to stay focused on one problem to avoid being overwhelmed by the vastness of AI tools. This hands-on experience is crucial for "future-proofing" oneself.

The Unprecedented Rate of AI Advancement

A significant misconception is that AI is only for "tech whiz kids." The speaker argues that curiosity and perseverance are sufficient to excel with AI. The term "prompt engineering" is dismissed as potentially overblown, as AI is becoming intuitive enough to understand natural language. The rate of AI development is described as "mind-blowing," surpassing even the iPhone's evolution. The speaker states that the "length of task and the power and potency of AI is doubling every seven months." This rapid progress means that past failures or current limitations of AI should not deter future exploration, as capabilities will significantly improve. The speaker urges trust in this progression, citing "undeniable" data.

The End of the Knowledge Economy and the Future of Work

The speaker foresees the end of the knowledge economy, where success was tied to learning and regurgitating information (e.g., accounting, law, radiology, coding). The future, according to the speaker, will prioritize creativity, problem-solving, and critical thinking. The ability to access and process information instantly renders traditional memorization less valuable. This shift is expected to lead to significant job losses, mirroring the layoffs seen in big tech companies like Meta and Amazon, and eventually impacting mainstream corporate America. The imperative is to be on the "right side of this incredible change."

The AI Value Chain Explained

The AI value chain is broken down into several key components:

• Infrastructure: This primarily involves GPUs (Graphics Processing Units), originally designed for video games but now crucial for AI. Companies like Nvidia are key players. The infrastructure is also "power hungry," driving investments in alternative energy sources by companies like Amazon and Microsoft to meet the demand for AI compute.
• Large Language Models (LLMs): These are the core of generative AI. The speaker draws a parallel to the early days of the internet (1994), where understanding its meaning was a challenge. Similarly, ChatGPT has brought AI to the masses, achieving 1 million users at an unprecedented speed. An LLM has an input (prompt), processing technology that analyzes vast data points, and an output (generated content). While similar to Google search in providing information, LLMs are multimodal and conversational. The race towards AGI (Artificial General Intelligence), where machines surpass human intelligence, is accelerating. The IQ score of the smartest LLM has reportedly increased from 96 to 136 in one year. Companies like Meta are developing open-source models like Llama, and new types of models, such as reasoning models that employ critical thinking, are emerging.
• Data: Data is the differentiator in AI applications. The speaker highlights legal battles where companies are suing over the use of their data to train LLMs (e.g., Reddit suing Anthropic, Disney and Universal suing Midjourney). The speaker reiterates the importance of personal data for their health bot and provides an example of using NYC open data (crime rates, air quality, building permits) to create an AI agent for real estate agents. For finance professionals, this translates to analyzing balance sheets, P&L statements, and cash flow statements to understand the "math equation of the business" and drive strategic decisions.
• Applications: This layer encompasses various forms of AI output, including text, video, image, voice, and data visualizations.
  • Text-to-Image: Tools like Midjourney have shown remarkable progression, with early versions producing flawed images (e.g., incorrect finger counts) and later versions achieving high fidelity. The speaker emphasizes that the problem to be solved dictates the application.
  • Text-to-Video: Google's V3 model is presented as a tool that can generate high-quality 4K video from prompts. While currently short, it's projected to evolve into full-length movies. The speaker demonstrates this by showing AI-generated videos for the presentation, highlighting the ease of creating visual content for communication.
  • Voice AI: While past voice assistants like Siri and Alexa were frustrating, current AI voice transcription accuracy is improving significantly, with potential applications in auditing and interviewing within the accounting industry.
  • Digital Twins: The ability to clone individuals and their voices raises new possibilities but also implications for deepfakes and fraud.
  • Coding: Tools like Cursor and CodiumAI are enabling complex application development with minimal prompts, leading to a dramatic drop in software development job postings and companies achieving significant revenue with smaller teams.

The Future of Finance and Strategic Decision-Making with AI

The speaker addresses the implications of AI for finance professionals, acknowledging that skills honed over careers may be automated. However, these skills can be combined with AI to propel careers forward. The key is to build something with AI, starting with personal projects.

The adoption curve of AI is moving from simple call-and-response (AI 1.0) to automation (AI 2.0) and now towards AI agents (AI 3.0). AI agents act as co-pilots, utilizing various tools to complete tasks.

Key applications for finance include:

• AI Agents in Operations: 75% of finance leaders expect AI agents to be routine in operations within three years.
• FP&A (Financial Planning & Analysis): AI can significantly improve forecasting beyond six months by modeling past data, similar to how the health bot identified patterns. This provides confidence for making decisions that unlock growth.
• Talking to Data: The ability to query and understand financial data through natural language interfaces is a crucial unlock. This allows for greater accessibility of business insights across the organization, enabling everyone to understand the "math equation" of the business.
• Business as a Math Equation: AI allows for the quantification of business processes (e.g., sales funnel conversion rates), enabling more strategic decision-making and driving change.

Conclusion and Call to Action

The speaker concludes by emphasizing that the future of knowledge-based functions lies in leveling up thinking and putting systems in place for easy data access. The job becomes less about solving the problem and more about what to do with the insights. The ultimate unlock is the ability to understand and leverage data to drive decisions and growth. The speaker encourages the audience to "do the work needed to future-proof themselves," acknowledging that while it's a "scary world," it's also a "powerful, exciting world." A QR code is provided for access to tools and resources discussed.