Before
An AI bookkeeping assistant for Korean freelancers that connects bank accounts, categorizes expenses, and prepares VAT filings.
After
B+An AI VAT assistant for Korean freelance designers earning ₩30M–₩100M that ingests a bank CSV plus KakaoTalk receipt photos and produces a Hometax-ready filing (not just a spreadsheet), replacing the ₩150k–₩300k accountant fee for the last mile.
Per-question score table
What's the strongest evidence you have that someone actually wants this?
Specific segment, income range, workflow, and current pain point clearly named.
What are your users doing right now to solve this problem, and what does that workaround cost them?
Concrete workflow, real spend, measurable time cost, named tools and prices.
Name the actual human who needs this most. What's their title and what consequence do they face?
Three real, compounding shifts — data access, model capability, regulatory tailwind — all named specifically.
What's the smallest possible version of this that someone would pay real money for this week?
Tight scope, real users, deliverable output, defers everything non-essential.
What surprised you when you actually watched someone use this without helping them?
Named channels, concrete script, manual delivery, built-in referral loop, watches real usage.
If the world looks meaningfully different in 3 years, does your product become more essential or less?
Real revenue target, named retention metric, tied to actual user behavior change.
Homework
Three concrete next steps before you run the next session.
- 1Name and DM 10 specific Korean freelance designers this week; watch 3 of them do their next VAT reconciliation over screen share and write down what surprised you about the categorization step.
- 2Rebuild V1 scope to output a filled-out Hometax form the user reviews and submits, not a spreadsheet handoff — own the last mile so the ₩150k accountant fee actually goes away.
- 3Pre-commit in writing: if month 6 hits 100 users at ₩19k but retention is under 40%, you raise price to ₩49k and re-test rather than discount; define the exact retention threshold that triggers a pricing change vs. a product pivot.
Keep using gstack on the web
/office-hours
- Run the full six-question workflow
- Get a refined V1 and grade
- Leave with concrete homework
- No terminal or local setup
All web skills
- 10 gstack skill runs every month
- Office Hours, CEO, Engineering, and Design
- Private results and saved history
- Follow-up emails
$19 / month
Go Pro
$19/month for 10 gstack skill runs across Office Hours, CEO, Engineering, and Design workflows—plus private results, saved history, and follow-up emails.
Liked your result? Share it.