Mentor Brief — March 2026
Budgeting that
actually sticks.
Artos removes the #1 reason people quit budgeting apps: manual data entry. Upload a screenshot or PDF — AI does the rest.
01 — The Problem
Students don't quit budgeting because they don't care. They quit because it's tedious.
"I'm an international student from Indonesia. Everything in Singapore is expensive. I used Money Manager — but between uni deadlines and CCA commitments, I'd only update it a few times a month. One-by-one transaction input is just too slow when you're busy."
There's a second, less-talked-about problem: outlier transactions. Paying a $3,000 semester fee or buying a new MacBook in an emergency — these aren't monthly spending habits, but they wreck your pie chart. Every app lumps them in. You either leave them out entirely (losing the record) or include them and your whole month's stats look wrong.
Too time-consuming
Manual entry per transaction kills consistency. Students give up after 2 weeks.
Outliers ruin stats
Tuition fees, medical bills, laptops — one large payment breaks the whole month's view.
Singapore banks are behind
MariBank, PayLah, GXS don't connect to budgeting apps via API. Screenshots are all you have.
Monthly-only is too late
Downloading a monthly PDF means you review spending after the damage is done.
02 — The Solution
Artos: semi-automatic budgeting with intelligent outlier handling.
Two inputs, one tap, done. Take a screenshot mid-month to stay on track. Upload the month-end PDF to consolidate. Claude AI reads both.
-
Screenshot analyser
Photograph your transaction screen any time. No CSV export needed. Works mid-month so you can catch overspending before the month ends.
-
PDF bulk import
Upload multiple months of e-statements in one go. Supports MariBank and DBS PayLah today; DBS, OCBC, UOB on the roadmap.
-
Unusual transaction flagging
Mark a payment as "Unusual" with one tap. It moves to a separate ledger and is excluded from your monthly stats and charts — but you still keep the record. No other consumer budgeting app does this.
-
Honest dashboard
Spending pie chart, category bars, daily averages, savings rate — all calculated on your real, outlier-free monthly spending.
03 — Competitive Landscape
No existing app solves the full problem.
Every competitor handles some of the pain — none handles all of it, especially not for Singapore's digital banking ecosystem.
| App | AI screenshot | PDF import | Outlier flag | SG banks |
|---|---|---|---|---|
| ✦ Artos | ✓ | ✓ | ✓ | ✓ |
| Money Manager | ✗ | ✗ | ✗ | ~ |
| Seedly | ✗ | ✗ | ✗ | ~ |
| YNAB | ✗ | ~ | ✗ | ✗ |
| Mint (shut down) | ✗ | ✗ | ✗ | ✗ |
✓ = supported | ~ = partial/limited | ✗ = not supported
04 — Market Opportunity
A growing, underserved segment at the intersection of fintech and AI.
Starting with Singapore-based students and young professionals who use digital banks — a segment that is actively growing as MariBank, GXS, and Trust Bank gain users.
05 — Why Me
I'm not building for a market. I'm building for myself two years ago.
Before university, I had a chronic online shopping problem — designer clothing, impulse purchases, no visibility on what I was spending. I didn't realise how bad it was until I started budgeting seriously two years ago. Seeing the numbers for the first time was genuinely shocking.
That habit shift changed how I make decisions. Now, before any purchase, I think about how it'll look on my end-of-month chart. I want my peers — especially international students navigating an expensive city on a student budget — to have that same moment of clarity, without the friction that made me almost give up three times.
I'm an international student from Indonesia, studying in Singapore. I'm my own target user. I feel every pain point I'm building against.
06 — Current Build
Working prototype. Live on the web today.
Built solo in the past few weeks. All core features are functional and deployed.
-
PDF + screenshot extraction
MariBank and DBS PayLah PDFs. Any Singapore bank screenshot. Claude AI reads both.
-
Unusual transaction system
One-click flag. Separate ledger. Excluded from charts and stats automatically.
-
Multi-account support
Track multiple bank accounts with individual balances, start dates, and opening funds.
-
Dashboard with real stats
Pie chart, spending by category, daily averages, savings rate, biggest expense.
-
Custom categories + bulk delete
Add your own categories, hide defaults, select multiple transactions to delete at once.
07 — Roadmap
12 weeks of focused development planned around real constraints.
08 — Business Model
Start free. Grow the user base. Monetise responsibly.
Freemium — Now to 2027
Core features free. Build habit, grow retention, reach product-market fit. Premium tier for power users: longer history, advanced AI insights, priority processing.
B2B Data — 2027+
Once meaningful user base is established: sell anonymised, aggregated spending insights to financial institutions, student-focused retailers, and insurance companies. Always opt-in, always anonymous.
Bank partnerships — Long term
Potential for white-label integration with digital banks (MariBank, GXS) who want to offer in-app budgeting without building it themselves.
09 — What I Need
Guidance, not funding — at this stage.
Feedback on the roadmap prioritisation. Am I building the right things in the right order? Should auth come before or after getting more user feedback?
Introductions. If you know students, NUS/NTU clubs, or fintech communities in Singapore who could be early testers — that would be the most valuable thing right now.
Validation on the outlier feature. Is this a real differentiator worth leading with, or should I talk about the AI extraction first?