What If This Were a Product?

When Project Delphy started working well, I did what most builders do: I started imagining users. Real users. People uploading their Vanguard statements at 7am, getting 10 AI analysts arguing about their tech allocation, then a CIO synthesizing the disagreement into an actual recommendation.

The vision felt clear. So I decided to take it seriously and actually stress-test it, the same way I would any business problem in my MSBA coursework. Not to validate what I hoped, but to find out what was actually true.

What follows is the full analysis: every path I explored, every number I ran, and where I landed.

The Initial Pitch to Myself

I started by writing the bull case. Not to be naive, but because any honest analysis has to take the opportunity seriously before poking holes in it.

The pitch wrote itself easily: there are 100 million retail investors in the US, most of whom have never received institutional-quality portfolio analysis. Bloomberg charges $24,000 per year. The gap between "nothing" and "Bloomberg" is enormous, and AI finally makes it possible to sit in that gap at a price point consumers can actually afford.

A good strategy consultant does not just write the pitch. They find out why the trap is a trap.

Path 1: Consumer SaaS

The most obvious route: a $9.99/month subscription, launch on ProductHunt, grow via word of mouth among the personal finance community on Reddit and Twitter. I modeled this first.

The unit economics looked acceptable at average usage. Then I hit the first real obstacle: the legal question.

The Investment Advisers Act of 1940 defines investment advice broadly: any person who, for compensation, advises others on the purchase or sale of securities. A personalized, for-pay AI tool that tells you whether to rebalance your portfolio almost certainly qualifies, regardless of a "not financial advice" disclaimer in the footer.

To cover $60,000 in compliance costs at $9.99/month with 70% gross margin, I need roughly 900 paying subscribers before turning a dollar of profit on compliance alone. That is before server costs, customer acquisition, or any product development.

I did not stop here though. A $60,000 compliance cost is manageable at sufficient scale. The harder question is whether the scale is achievable given the competitive landscape.

The Competitive Problem

I mapped out who would actually be competing for the same users.

The distribution problem is the real killer. Fidelity and Schwab have users who already trust them with their financial data. For a new product to win, it has to be meaningfully better in a way the user can articulate in under 30 seconds and worth the switching cost of uploading statements to a startup they have never heard of.

Project Delphy's Delphi architecture is a real differentiator. But it is an architectural one: it can be replicated by any team with API access in a matter of weeks. There is no proprietary model, no exclusive data feed, no network effect that compounds with scale. Technical differentiation without a defensible moat is a head start, not a competitive advantage.

Path 2: Selling to Financial Advisors

When consumer felt too hard, I pivoted the model. What if the customer was not the retail investor, but the financial advisor who serves them?

The pitch here is stronger in several ways. Financial advisors already pay for research tools. They have higher willingness to pay ($99–299/month instead of $9.99). And critically, the advisor carries the fiduciary responsibility, which sidesteps much of the retail regulatory complexity.

The B2B path is more viable on paper. But it requires building a company, not a product. A sales team. A customer success function. An enterprise security posture. None of that work is interesting AI systems work, and none of it can be done on the side.

The honest answer: pursuing B2B would require raising money, hiring people, and committing 2+ years before seeing meaningful traction. That is a legitimate business decision for a venture-backed startup. It is not the right path for a solo MSBA student project with a full course load.

Path 3: Platform Licensing

My third angle: do not compete with Fidelity and Schwab. Sell to them.

The pitch is that Project Delphy's architecture is a genuinely novel approach to AI financial analysis. A large platform with existing users, compliance infrastructure, and a distribution moat could integrate this as a differentiated feature. They absorb the regulatory cost; the technology becomes their product.

I spent time thinking through how realistic this was. The blockers are not the technology. They are structural:

Platform licensing is theoretically attractive and practically inaccessible. Not impossible, but it requires a network and a track record that does not yet exist.

What the Data Actually Says

After running all three paths, the picture is clearer than I expected.

The conclusion is not that commercialization is impossible. It is that it requires more than I can commit to at this stage: either institutional funding for the compliance infrastructure, or a co-founder with a BD background and the time to build the B2B sales motion.

Neither condition is met. So the question becomes: what is the right use of this work given that constraint?

The Actual Decision

Project Delphy has real value as a demonstration of systems thinking, ML intuition applied to AI orchestration, and full-stack product development. These are things I want a recruiter or PM to see clearly when they look at my work.

Shipping this as an open-source project achieves that. It shows the technical work without requiring me to pretend I have a commercial path that does not yet exist. And if a platform or advisor tool ever does want to license the architecture, an established open-source repo with a documented analysis is actually a better starting position than a barely-launched consumer product.

The decision to open-source is not a consolation prize. It is the correct allocation of this asset given the current constraints.