Counterfactual - Portfolio Dependency Engine for Investors

Note: This analysis is based entirely on the self-reported project description provided by the authors. No external verification, traction data, revenue figures, or customer information are available.

The description states that Counterfactual turns a personal portfolio into an “interactive Portfolio X-ray,” which shows how investments depend on underlying factors such as companies, sectors, countries, macro conditions, and themes. It uses a normalized portfolio layer to support integration with brokers like Alpaca.

Six core workflows are described:

• Overview
• Dependencies (graph and map)
• Scenario Lab
• The Wire (curated news/scenarios)
• Simulation
• Guide

The system is said to propagate dependencies through a deterministic engine that accounts for relationship strength, confidence, freshness, transmission lag, cycles, and converging paths.

Claim: Counterfactual maps hidden dependencies in portfolios.

Evidence: Described as a "Portfolio X-ray" with six workflows including graphing, mapping, scenario modeling, and simulation.

Inference: The product is designed to help investors understand what their portfolio actually depends on — not just what they own.

The project positions itself as a tool that moves beyond traditional portfolio tracking ("what do I own?") to deeper risk analysis ("what does everything I own depend on?"). It claims to offer a "risk-desk view" of assumptions connecting investments, and to provide evidence-based action proposals.

It also states that it doesn’t replace judgment but enhances it with better maps and sharper questions.

Claim: Counterfactual helps investors move from ownership tracking to dependency mapping.

Evidence: The tagline: “Bloomberg tells you what happened. Counterfactual maps what your portfolio depends on, models what comes next, so you act with evidence.”

Inference: This is a shift from passive reporting to active risk modeling.

The description implies that the target customer is an individual investor managing their own money — not institutional or retail investors specifically named. The team notes that they built it for “a regular investor” rather than quants or traders.

Claim: The tool targets individual investors.

Evidence: “The goal stays the same: make serious portfolio reasoning useful to anyone managing their own money.”

Inference: Not explicitly defined by segment, but implied to be self-directed investors seeking deeper insight into portfolio risk.

No business model or pricing information is provided in the description. The project is presented as a hackathon submission and does not mention monetization strategies, subscription tiers, or any commercial framework.

Claim: No stated business model or pricing.

Evidence: Not evidenced.

Inference: Likely early-stage prototype with no revenue path described.

The product is built using:

• Next.js, React, TypeScript
• React Flow for graphing
• Recharts for visualization
• Alpaca APIs for broker integration
• OpenAI GPT-5.6 and Codex for development assistance

It includes a knowledge engine with stages: curated evidence → AI-assisted extraction → RL-assisted scoring → validation → versioned dependency graph.

The system is said to be deterministic, with shared analytical contracts across interfaces to ensure consistency.

Claim: Technical stack and architecture are described.

Evidence: Listed technologies and workflow stages.

Inference: The use of AI tools (especially Codex) suggests a rapid prototyping approach; however, no production-ready infrastructure or scalability details are given.

There is no evidence of traction, customers, or product adoption. The project was submitted to an OpenAI hackathon and built in one week. It uses Alpaca’s paper environment for demonstration only.

Claim: No traction or maturity signals.

Evidence: Not evidenced.

Inference: Likely pre-product-market fit stage; no real-world usage data.

The description does not mention competitors or direct market positioning. The author references Bloomberg as a comparison point, but no other tools or platforms are named.

Claim: No competitive context provided.

Evidence: Not evidenced.

Inference: The tool may overlap with portfolio risk analysis or dependency mapping tools, but no specific competitive landscape is described.

• Unproven user trust: There is no evidence that users would actually act on the dependency maps or scenario models.
• AI over-reliance: While AI was used heavily in development, there is no indication of how human oversight is maintained in actual use.
• Limited scope: Only Alpaca integration is mentioned; broader broker support is stated as future work.
• No validation mechanism for AI outputs: The system relies on AI for relationship extraction and scoring but lacks details on how accuracy or bias are managed.

Claim: Risks related to user adoption, AI trustworthiness, and limited functionality.

Evidence: Absence of user feedback, lack of production-ready features, reliance on AI without validation layers.

Inference: These could hinder long-term viability if not addressed in later development.

1. What specific types of dependencies (e.g., sector, geography, macro) are currently modeled and how are they validated?
2. How does the system handle edge cases like unmapped holdings or conflicting data sources?
3. Has there been any user testing or feedback on the clarity and usefulness of the dependency graphs and scenario simulations?
4. What is the plan for expanding beyond Alpaca to other brokers?
5. How will the tool ensure that AI-generated insights are not misinterpreted by users?
6. Are there plans to incorporate licensed historical data for calibration of simulations?