Suite PYMES — the back office inside the shop
In-house product
Consumer retail with in-house delivery
2026
The problem
A consumer retail business with its own delivery decides every day what to restock, which supplier to buy from and what to charge for delivery. E-commerce platforms solve the shop window; everything else is left out.
Stock, supplier costs, driver settlements and real profitability end up in a spreadsheet someone updates when they can, and in mental arithmetic that is almost always optimistic.
What was built
A system where selling and running the business are the same thing. The order comes in through the store, decrements stock atomically, is charged through MercadoPago, issues an electronic invoice automatically on payment, is assigned to a driver and lands on an accounting dashboard.
Around it, the support stack: WhatsApp wired into a unified inbox and an AI agent answering from the live catalogue and current prices.
The details that make the difference
The margin that was half lying
When the customer paid for delivery, that amount counted as income but the payment to the driver was not deducted as a cost. The problem was documented with figures, both accounting options were laid out with their effect on the percentage, an accountant was consulted, and only then did the formula change. The reported margin went down, and became the real one.
Free delivery that does not erase the cost
When the cart passes the threshold, the customer sees “free delivery”, but the system still records what that trip costs — to pay the driver, and so the dashboard shows what waiving it was worth.
The business day, not the server’s
Reports group sales by the shop’s own time zone, derived automatically from the pin the owner drops on the map. A 22:00 order counts on its own day, not on the server clock’s next one.
Help that does not go stale
Instead of a manual full of screenshots that age at the first change, interactive tours that point at the real screen and remember where each user left off.
An AI budget that holds
The system reserves the spend before calling the model, never after. If there is not enough, it spends no tokens at all and hands the conversation to a person.
How it was verified
Verification was a self-run audit in seven phases across 92 acceptance criteria: documentation inventory, contrast between what was documented and what the code actually does, currency of every dependency, test plan, automated end-to-end execution with Playwright, security review and verified remediation.
It surfaced faults that never show up in a demo: a collaborator could mint themselves an administrator account; the brand-colour field allowed code injection across every public page; the catalogue could be manipulated so the AI agent quoted false prices; and under concurrent orders the AI spend cap was exceeded by 150%. Every fix shipped with its regression test.
Nothing is marked “working” without executed evidence. What was not tested is delivered as unverified, never as approved.
- Next.js 16
- TypeScript
- PostgreSQL
- Prisma
- Tailwind CSS
- Docker
- MercadoPago
- ARCA/AFIP
- Google Maps
- n8n
- Chatwoot
- Playwright
The hard question first: did the business make money last month, or does it look like it did?
If the answer is uncomfortable, there is a fifteen-minute conversation worth having. The system is shown running on a demonstration business, and the way the work happens today gets mapped.