Automated Manager
Recorded activity becomes priorities, cost reviews and decisions. Evidence, calculations and explicit assumptions stay attached to the recommendation.
Independent product case studyBatam, Indonesia · 2025–present
INDOWYSS
Building an operating ecosystem for small F&B businesses.
Daily records are the starting point, not the end product. INDOWYSS connects an automated Manager, receipt-based promotion and coordinated purchasing around the same business operations.
01 / The product
Working screens. Connected workflows.
Screens captured from the application during local QA. UI labels are in Indonesian; the descriptions below explain the purpose in English. These are not customer revenue or rollout results.
Automated Manager
From dashboard totals to a next action.
Daily issues and longer-term business plans have separate places. Each decision can show its recorded evidence and open the relevant operational workflow.
01 / 06
Decision support
Cross Promo
A participating outlet's promotion appears on another outlet's customer receipt. Owners control the offer and its conditions.
Collective purchasing & supplier matching
Compatible needs meet supplier offers and minimum quantities. Buyers review quality, price and fulfilment terms before a final commitment.
Modern POS & daily operations
Transactions, receipts, shifts, expenses, closing, stock and menu costs provide the records that make the ecosystem useful.
02 / Management thinking
A recommendation needs more than a number.
The Manager follows a practical sequence: recorded evidence, a calculation or scenario when needed, then a next action. Missing costs limit the conclusion instead of being filled in with invented figures.
Example workflow
An ingredient purchase cost changes.
- Capture the change
Record the purchase and review the affected menu-cost basis.
- Compare the options
Calculate the effect using known costs, sales and stated assumptions.
- Keep the owner in control
Price changes and other consequential actions require authorised approval.
- Check what happened afterwards
Compare later records with the earlier period; do not treat correlation as proof.
Simulate before spending
Staffing and equipment scenarios compare additional sales contribution with additional costs. Equipment payback is conditional on the assumed benefit actually occurring.
Evaluate without overstating
Before-and-after reviews keep menu costs and extra trial costs visible. A better period is not automatically net profit or proof that one action caused the improvement.
03 / The business network
Operations can create opportunities beyond one outlet.
Cross Promo: a receipt becomes a route to another business.
A customer who shops at outlet A can see a promotion from participating outlet B on the receipt. Outlet B sets the terms and the customer follows the claim path. Visibility is an opportunity, not a promise of conversion.
- 01Purchase at outlet AOrdinary transaction
- 02Receipt carries B's offerOwner-defined conditions
- 03Customer visits outlet BVoucher / QR claim path
Payment, refund and settlement states are modelled through sandbox and partner-adapter boundaries. Production money movement is a separate activation step.
04 / INDOWYSS OS & AI design
The foundation is a data-to-decision system.
The OS approach defines how records should become trustworthy context, decisions and follow-up. Current Manager calculations and rules work from operational records; external AI is not the source of financial truth.
- 01Capture
Transactions, closing, purchases and costs.
- 02Validate
Inputs, permissions and outlet scope.
- 03Check quality
Completeness and usable evidence.
- 04Calculate
Backend figures and explicit assumptions.
- 05Prioritise & explain
Evidence, impact and a next action.
- 06Review & evaluate
Human approval, feedback and later records.
AI helps explain; the backend calculates; the owner decides. Structured outlet context and outcome records can strengthen subsequent analysis. They do not mean an external model automatically retrains itself or that every product feature already uses one unified OS pipeline.
05 / Quality & delivery
Test the failure paths, not only the happy path.
Validation work has included unit checks, real local PostgreSQL/API testing, production builds and browser QA across desktop and mobile. These describe development practice, not a claim that all public-production risks are resolved.
Concurrent closing
Review transactions arriving around shift closure and server-side financial totals.
Repeated refunds & sync
Check duplicate requests, stock return and already-refunded or voided records.
Queue unavailable
Keep closing independent of immediate queue availability and preserve follow-up work.
Authorisation & data boundaries
Scope records by outlet and require the appropriate role for consequential changes.
Current delivery status
- Working local application
- Core operations, calculation/rule-based Manager, simulations, evaluation records and network workflows have been built and locally validated.
- Restaurant beta
- Tables, orders, kitchen and staff workflows extend the existing operational records in a separate beta lane. This is not a claim of hotel-scale deployment.
- Separate activation steps
- Paid external AI, actual WhatsApp delivery and production payment processing are not active. Commercial rollout and measured customer outcomes are not claimed here.
06 / My contribution
From a field problem to a working product.
I am Wisnu Falah, the independent founder and product builder behind INDOWYSS. The project started with manual records and AI-assisted analysis for small F&B operators in Batam in 2025, then evolved into a structured software product.
- Product direction
- Identify frontline problems, define priorities and turn them into requirements and acceptance criteria.
- Workflow & data design
- Define the records, roles, approvals, state transitions, failure paths and evidence needed for a decision.
- AI-assisted implementation
- Build and iterate frontend, backend and database workflows, with documentation and scoped validation.
- Owner experience
- Refine navigation, plain-language reports and progressive detail so a non-technical owner can act.
- Testing & follow-up
- Review scenarios and regressions, protect demonstration environments and connect actions with later records.
Development stack used TypeScript · Next.js · NestJS · PostgreSQL · Prisma · Git
Practical background Family retail operations · Part-time F&B service · Sales · English teaching
This is a product-development portfolio, not previously held hotel departmental-management employment. Private source code, credentials and customer records are excluded.
Documents & contact
A concise CV. A deeper project discussion.
CVPrevious standalone version · 1 page · PDFINDOWYSS project briefPrevious standalone version · 2 pages · PDF
Wisnu Falahfalahwisnu711@gmail.com+62 819 1024 9645