AI Engineer

Why Off-the-Shelf AI Doesn't Understand Money — Udi Menkes, Intuit

Published 2026-07-29 · Duration 19:50

Summary

Off-the-shelf Large Language Models (LLMs) often provide fluent but inaccurate financial advice—a phenomenon termed the 'fluent bluff'—because they lack real-world experience. The core argument is that mere context is not sufficient; true intelligence requires grounding in verifiable business outcomes. To solve this, Intuit builds systems that transform raw business data (e.g., General Ledger, P&L) into millions of State-Action-Outcome records. These structured datasets are used to train models that predict the actual impact of actions, leading to 'outcome-driven finance' and outperforming frontier LLMs in real-world decision-making.

Download summary

Key takeaways

  1. The Fluent Bluff: Why Off-the-Shelf Models Fail at Money 4:08

    LLMs trained on general internet data (blogs, books) can give confident advice that is fundamentally wrong for specific business constraints. Examples include recommending a negative cash flow business acquire another property or raising prices when it risks losing all revenue.

  2. The Solution: Grounding in State-Action-Outcome Data

    Instead of relying on general knowledge, the system must be grounded by creating structured records detailing what a company *did* (Action) given its current financial status (State), and what the verifiable result was (Outcome). This moves advice from theoretical possibility to proven reality.

  3. Measuring Impact: Beyond Simple Difference 15:45

    Determining the true impact of an action requires advanced statistical methods like Conditional Average Treatment Error (CATE) to account for confounding variables, such as a company's natural propensity for success. Failing to do this leads to inaccurate conclusions about causality.

Technical details

  • Data Modeling and Training

    The system creates millions of vectors by combining financial data from sources like QuickBooks (General Ledger, P&L, Cash Flow statements) into State-Action-Outcome records. This structured data is used to train a Reinforcement Learning (RL) model that predicts optimal actions.

  • Model Architecture

    The process involves three steps: 1) Defining the business state; 2) Deriving potential actions from the state; and 3) Using an RL-trained model to predict outcomes, followed by training a final LLM layer to generate evidence-based advice.

  • Performance Benchmarking

    The grounded system was tested head-to-head against leading frontier models and outperformed them because of its access to verifiable outcome data, not just model size or general knowledge.

Mentioned resources

  • Princeton Study (Research/Study)
  • QuickBooks, Turboax, Credit Karma, Mailchimp (Product Data Sources)

Channel & topics

Watch on YouTube · Back to latest

This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.