Four attempts. One that finally worked.

It started in thrift stores

Jason grew up watching his parents clip coupons and shop at Goodwill — not because they enjoyed it, but because the math demanded it. They worked hard, earned decent wages, and still ended up with nothing saved by retirement. No one had ever sat down with them to run the numbers. No one had ever explained that a small pension change at age 62 versus 65 could mean $180,000 over a lifetime.

That gap — between people who could afford a financial planner and people who couldn't — never left him.

Attempt one: become the CFP

Jason earned his CFP designation and went to work at a fee-only planning firm called Open Source Connections. The work was meaningful. He sat with real families, ran Monte Carlo simulations, modeled Social Security claiming strategies, and watched people light up when they finally understood their own situation.

But the minimum engagement was $2,500. The clients who most needed the analysis couldn't pay for it. And the ones who could often just wanted someone to validate what they'd already decided. Something felt backwards.

Attempt two: go solo and lower the price

He started his own one-person planning practice. Lower overhead meant lower prices — he could get a comprehensive plan down to $1,200. Still too high for the families he cared most about. And the work was unsustainable: every analysis was 30–40 hours of custom modeling, coordination scheduling, and report writing. There was no way to serve hundreds of people at that rate.

He started putting the planning model into spreadsheets. Refining it. Documenting every assumption. What he didn't realize was he was building a deterministic engine — he just didn't have software to run it.

Attempt three: myFinancialAnswers

The third attempt had a name, a logo, and a domain. He hired a developer, built an early web app, and launched myFinancialAnswers — a tool that would run a subset of the planning engine online. He put $40,000 on credit cards. His wife was pregnant with their first child.

The product worked well enough. The business didn't. Acquiring users was harder than he expected. The app answered questions but didn't guide people through the right questions. Without the conversational layer that a live CFP provides, users filled in the fields wrong, got confusing outputs, and bounced. The credit card debt outlasted the startup.

He shut it down and went back to employment. But the spreadsheet model kept getting better.

Attempt four: Broadtree

A few years later, he tried again with a B2B angle — a product called Broadtree aimed at HR departments and credit unions. The insight was sound: employers already paid for benefits, so packaging financial planning as a benefit solved the pricing problem. He got close. Pilots were promising. But closing enterprise deals on a founder's timeline, with no sales team and a product that still required significant setup per client, stalled him out.

By this point he had a cleaner framework for why each attempt failed. Attempt one and two: the bottleneck was his own time — financial planning required a human in the loop. Attempt three: the UI couldn't replace the guidance a planner provides. Attempt four: the sales cycle was too long for a solo founder.

What he needed was something that could replace the conversational guidance without requiring a human. He just didn't have the technology.

The ChatGPT moment

In late 2022, he sat down with ChatGPT and typed in the first few questions from his onboarding script. It asked follow-up questions. It rephrased when a user was confused. It did exactly what the missing conversational layer needed to do.

He was careful not to let the AI generate financial numbers — he'd seen enough hallucinated calculations to know that was a dead end. But as an interface layer, sitting between a user and a validated deterministic engine? That was different. The calculations stayed deterministic, auditable, and testable. The AI just helped people give the engine the right inputs and understand what came out.

He rebuilt the spreadsheet model in TypeScript. Every formula. Every edge case he'd encountered across fifteen years of planning. He wrote a test suite that verified every calculation against hand-validated known answers. Then he wired the AI layer in front of it.

AI Financial Plan

The platform you're using today is attempt five. It solves the problems that killed the first four: the calculations are automated (no human bottleneck), the AI guides collection without requiring a planner, the price is accessible, and the B2B channel makes employer distribution tractable.

It is not a chatbot guessing at your financial future. It is a deterministic planning engine — the same 8-module framework used by CFPs, now available to anyone with a browser.

The families who couldn't pay $2,500 deserve to know whether their retirement is on track. They deserve a Social Security claiming strategy. They deserve someone to run the long-term care numbers before it's too late to act on them.

That's what this is for.

A note on what this is — and isn't

This platform provides financial education and analysis, not investment advice. The calculations are the same ones a CFP would run. The AI explains them in plain language. But the outputs are analysis — they help you understand your situation and the tradeoffs in front of you. For implementation decisions, consult a fee-only financial planner, tax professional, or estate attorney as appropriate. We tell you that not to disclaim responsibility, but because it's genuinely true that a good advisor earns their fee when you're ready to act on analysis like this.