Context
ArcOunt is a personal financial journey companion — unifying budgeting, goal planning, and financial education into one product, with an architecture designed to work beyond a single reference market from day one.
Problem
People who want to take their financial lives seriously are underserved at both ends of the market: general accounting software assumes a level of financial literacy most people don't have, while consumer savings apps are good at executing a plan you already know you want — neither builds the plan with you or teaches along the way.
Opportunity
The gap sits between "teach me" and "do it for me" — nothing in the reference market combines financial education and budgeting execution into one companion that meets a user wherever they are on that spectrum and moves them along it over time.
Research
Personas were built from specific, real user situations rather than a generic "saver" archetype, and I mapped the competitive landscape by category — accounting software, consumer savings apps, financial-education content — rather than by individual product, to keep the analysis focused on the structural gap rather than any one competitor's feature list.
Decision-Making
A set of explicit product principles was written before any scoping call and used as the decision filter for every trade-off that followed — including which of the product's pillars to build first, and how much of the financial-education layer to gate behind onboarding versus surface immediately.
Prioritization
An 11-entity data model and a phased roadmap sequenced the AI-agent product layer behind explicit consent and no-fabrication guardrails, built deliberately ahead of user-facing polish — a companion giving financial guidance has to earn trust structurally before it earns it visually.
Execution
The strategy translated into 12+ iterative prototypes and a working interactive demo, each iteration testing a specific hypothesis about the onboarding-to-education handoff rather than a general "make it better" pass.
Collaboration
As sole product owner, I used an AI assistant as a structured drafting and stress-testing partner — feeding it the persona work, principles, and data model to accelerate documentation and pressure-test edge cases — while every scope call and trade-off decision stayed mine.
Trade-offs
The clearest trade-off was sequencing: building the AI-agent layer's guardrails before the interface it would eventually sit behind meant slower visible progress early, in exchange for not retrofitting consent and fabrication safeguards into a product people were already trusting with financial data.
Outcome
A live interactive prototype is linked above. The full PRD, data model, and roadmap are product-strategy documents I keep private at this stage and share directly on request rather than publish here in full.
Reflection
Building the guardrails before the interface confirmed something I'd only had a hypothesis about going in: trust-sensitive products need their trust mechanics designed structurally, not bolted on once the UI looks finished.
Lessons Learned
AI collaboration is most valuable when the human supplies the framing, principles, and trade-off decisions, and the AI accelerates structured drafting and iteration. The quality of this work came from the personas, principles, and compliance strategy behind it — none of which an AI would originate unprompted.
Future Improvements
The next real test isn't another prototype iteration — it's structured user interviews against the financial-education hypothesis specifically, to find out whether teaching-while-budgeting actually changes behavior or just adds friction.
My Contribution
Sole product owner, start to finish: problem framing, persona development, scope definition, the full data model, and roadmap sequencing were mine. AI accelerated structured drafting and prototyping under that direction.