PRODUCT JUDGMENT
Define the problem, journeys and principles before asking AI to implement them.PRODUCT · ENGINEERING · RESPONSIBLE AI
Staying on top of my financial life had started to feel like another job.
I was comfortable with my finances and the technology around them. What I did not have was the time to keep checking different apps, maintaining spreadsheets and piecing together the full picture. Valenso began with the belief that the product should do more of that work for me.
01 · WHY VALENSO
I needed less financial admin.
That problem was not mine alone. When I tested it with friends and family, the same pattern kept coming back: people who understood their money were still spending too much time checking apps, maintaining spreadsheets and reconstructing the full picture.
Tools like Monarch and Rocket Money bring personal finances into one place. I wanted Valenso to go further: an AI-first financial manager that organizes the work, understands complete cash flow and points to what deserves attention—while explaining itself and keeping the user in control.
WHY NOW
Agentic engineering made this the right moment.
I had wanted to solve this problem for years. What changed was the arrival of capable AI agents that could help turn a carefully defined product vision into working software—fast enough to explore the whole system, not just prototype the surface.
This was where I wanted to bring my product, engineering and AI background together: use agents as implementation leverage, keep the important product and architecture decisions human-owned, and build Valenso in a way that could solve the problem for me first while creating a secure foundation that can scale to others.
AGENTIC LEVERAGE
Use agents to accelerate research, implementation, testing and iteration across the stack.ENGINEERING DISCIPLINE
Own the architecture, security boundaries, verification and production decisions.02 · THE PRODUCT
A clearer, more confident way to stay on top of your money.
Valenso connects bank accounts and credit cards, accepts statement uploads and organizes transactions into one traceable cash-flow story: what came in, where it went, what was saved and what changed in spendable cash.
The product is designed to be honest when it knows and clear when it does not. Every automated category has a reason, corrections are preserved, possible duplicates are surfaced, and every reported figure can be traced back to the transactions behind it.
THE GUIDING PRINCIPLE
A simple test for every decision.
Kanso is the practice of removing what is unnecessary so what matters becomes easier to see. I use it as a simple internal test for every product and design decision.
Does this remove financial clutter and make what matters easier to see?
DESIGN
Kanso in design
Calm, restrained and focused on what matters.LANGUAGE
Clarity in language
Understandable without simplification becoming vague.DATA
Trust in data
Stable numbers with explanations you can inspect.AI
Control in AI
Useful assistance without taking agency away.FROM WHY TO HOW
AI accelerated the build. Experience gave it a foundation.
My experience building financial products for a Fortune 500 company in a regulated environment shaped how I approached Valenso. Agentic engineering let me move faster across product, design and code. The foundation—clear customer journeys, accurate financial logic, security boundaries and accountable decisions—came from years of building products where trust matters.
PRODUCT
Start with the real problem
Turn a personal frustration into clear user journeys, priorities and measures of success.
ENGINEERING
Make the numbers dependable
Keep financial logic precise, traceable and protected as the product grows.
AGENTIC ENGINEERING
Use AI to expand the build
Give agents clear context, review their work, and use them across research, design, coding, testing and documentation.
SECURITY & GOVERNANCE
Set non-negotiables early
Decide what data can move, what AI can see and where people stay in control before shipping features.
03 · THE BUILD
Agentic did not mean improvisational.
The way I built Valenso was as important as what I built. AI increased the pace; a simple operating loop kept the work grounded.
01
Make intent explicit
Before asking an agent to build, I captured the user problem, the behavior I wanted, the boundaries it could not cross and what success would look like.
Problem · principles · acceptance criteria
02
Build in verifiable slices
Agents helped turn each decision into working software, tests and documentation. I reviewed the result as a product owner and an engineer before it moved forward.
Working software · automated checks · owner review
03
Use Valenso, then re-plan
Real use exposed the next valuable problem more reliably than a fixed roadmap. I changed priorities when the evidence changed.
Daily use · decision log · production feedback
AI shortened the distance from decision to working software. It did not remove the need for judgment, accountability or care.
04 · ARCHITECTURE
Keep the financial truth stable, even when the technology changes.
Valenso keeps its financial logic separate from the companies that supply bank data or AI. That makes those providers replaceable, keeps important decisions testable and puts a clear boundary around sensitive information.
EXPERIENCE
What the user sees
Dashboard · transactions · guided workflowsDOMAIN CORE
Financial logic
Money · reporting · matching · user choicesDECISION SYSTEM
Explainable rules
Ordered matching · clear reasons · reversible choicesPRIVATE DATA
Encrypted financial data
One user at a time · protected fields · audit historyBANK CONNECTIONS
Replaceable providers
Secure sync · common format · duplicate handlingCONTROLLED AI PATH
One route to AI
Approved fields · redaction · cost limits · activity logOnly approved, de-identified information can leave Valenso
Change providers without a rewrite
Bank-data and AI providers connect at the edge; Valenso’s financial logic stays independent.
Keep AI out of financial truth
Money, matching, reporting and user choices live in tested code. AI can suggest, but it never becomes the source of record.
Make releases reversible
Small database changes, feature switches and separate environments make it safer to learn and correct course.
05 · SECURITY & GOVERNANCE
Trust is built into the product—not added at the end.
Financial data raises the stakes. I designed Valenso so outside AI providers are treated as untrusted: the product decides exactly what may leave, records why it left and keeps sensitive information inside.
THE GOLDEN RULE
No personal or sensitive financial data goes to the model.
Information is blocked by default. It can be sent only after it is explicitly approved and passes checks for personal or payment data.Send the minimum
Models receive derived tokens and bounded context—not raw merchants, amounts, balances, account masks or transaction descriptions.
Use one controlled route
Every AI request follows the same guarded path. Tests prevent a new feature from quietly bypassing it.
Record what leaves
Valenso records the purpose and shape of every AI request without logging the sensitive value itself.
Control cost
AI usage is measured and capped; bank-provider usage is recorded so the cost of running the product stays visible.
Keep people in control
AI answers are suggestions. The user confirms the durable category, tag or rule, and can always see why it happened.
The wider security posture
Encrypted sensitive fields, strict per-user queries, MFA-gated account linking, signed provider webhooks, content security policy, secret scanning and isolated environments reinforce the AI boundary.
This is an engineered risk posture, not a claim of certification. Controls are documented, tested and designed to evolve as the product and regulatory context change.
CONTINUING THE BUILD
Valenso is a living product—and this is the record of how it is taking shape.
I continue to refine the product through real use, documenting the decisions, tradeoffs and lessons along the way. If you are curious about a part of the build, I am always happy to compare notes.
Ask Pranav about Valenso