Amela
RetailFlow Inc. — New Platform Launch Story
ℹ︎ For confidentiality, company names, project names and system images have been replaced with pseudonyms and mock-ups.

Ordering work,Built in-house from scratch to automate it.

StockLink is a distribution back-office automation platform that automates ordering based on Cloud POS inventory across multiple stores. From inventory checks to ordering and arrival matching, it consolidates everything into a single platform.

"Inventory data integration, ordering, external-system integration and operational monitoring" — the story of launching a new platform, built in-house from scratch in 8 months, that satisfies all four axes a distribution operation needs at once.

See the build story Try the demo See overview
Ordering workload
-85%
Reduction from manual transcription
Auto-ordering
24hrs unattended
RPA scheduled-batch operation
Inventory transcription errors
0cases
Auto-calculated from POS data
External integration
APINot needed (RPA)
Reproduce manual work via screen operation
admin.stocklink.example.jp/orders
6
To reorder
4
Auto-ordering
3
Arriving
Order suggestions RPA Auto-Order
Milk 1L×24
Sliced Bread (6)×30
Chicken Breast 2kgOrdered
01
CLIENT
About the client

RetailFlow Inc., one of the country's leading players driving retail and distribution.

RetailFlow Inc. is a retail and distribution company operating stores across multiple areas. Continuously sourcing goods from many suppliers in step with daily sales, the precision and speed of its procurement operations are the lifeline of the business.

While store Cloud POS systems accumulated rich inventory and sales data, ordering to suppliers stayed fragmented across separate portals and EDIs and dependent on manual work. To resolve this "the data exists but isn't connected" situation, the company chose to build an ordering platform in-house from the ground up.

Connected stores
4Locations
SKUs managed
1,200+
Integrated suppliers
Multiplecompanies
Business Flow — Auto-ordering driven by inventory data
1
STEP
Auto-fetch POS inventory
Collect inventory and sales data from Cloud POS
2
STEP
Auto-calculate order quantity
Extract reorder SKUs from thresholds and demand
3
STEP
Auto-order via RPA
Order to supplier portals and EDIs via screen operation
4
STEP
Auto-match arrivals
Match orders with arrivals and reflect into inventory
02
CHALLENGE
The design challenges to solve

The four structural requirements demanded of the platform.

None of these were questions of technology choice — they were business requirements tied directly to procurement cost, inventory accuracy and frontline workload. Building anew from scratch demanded a design that "makes them all hold true at once."

Inventory data stays locked at the frontline
Keep checking and transcribing POS inventory by hand, and both accuracy and speed hit a ceiling,Stockouts and overstock becomes the norm.
Ordering is fragmented per supplier
Order manually to each of many suppliers one by one, andEffort and errors swells, and the work becomes dependent on a single person.
External systems have no API
Leave supplier portals and EDIs without APIs untouched, andStalls just short of automation, human intervention never goes away.
People tied to routine processing
Get the operational design of late-night/early-morning routine processing wrong, andLabor cost and delays builds up, and round-the-clock operation never holds.
03
ARCHITECTURE
The architecture we built together

Turning ordering into a "platform that never stops" across four domains.

From capturing inventory data to ordering, external-system integration and operational monitoring — together with RetailFlow Inc. we built, in-house and from scratch, the four domains a distribution back-office needs. Even for partners without APIs, RPA bridges the gap, delivering a platform that keeps running unattended.

Amela EXECUTION
8months
Duration
32person-months
Total effort
Domain ①
Inventory Data Integration
Challenge
A need to break away from manually checking and transcribing store POS inventory.
Solution
Auto-fetch inventory and sales from Cloud POS
Make the captured data the starting point for order calculation.
Domain ②
Ordering
Challenge
A need to cut the effort and errors of manual ordering across many suppliers.
Solution
Auto-calculate reorder SKUs and quantities, processed in bulk
To multiple suppliers in one click.
Domain ③
External System Integration
Challenge
A need to auto-integrate even with API-less supplier portals and EDIs.
Solution
RPA automatically reproduces human screen operations
CSV and EDI formats are converted and integrated too.
Domain ④
Operations & Monitoring
Challenge
A need to detach late-night and early-morning routine processing from manual labor.
Solution
Unattended operation via scheduled batches, with results auto-notified
Monitor execution via Teams and email.
FRONTEND
React Ant Design React Query i18next
BACKEND
NestJS TypeScript JWT TypeORM Redis
AUTO & INFRA
Puppeteer (RPA) ⚡Scheduler / Worker PCloud POS integration ESupplier EDI / CSV
04
IMPACT
The value the platform creates

RetailFlow Inc.'s "procurement that never stops" accelerates.

This platform continuously delivers not technical metrics but business outcomes — procurement efficiency, inventory accuracy and how the frontline works.

A platform that runs automatically from inventory to ordering
Everything from inventory visibility to order execution keeps running without interruption.
Prevents stockouts and overstock
Order quantities are auto-calculated to match demand, optimizing inventory.
A design that maximizes integration even without APIs
RPA brings external systems on board, expanding the scope of automation.
People can focus on judgment work
Routine tasks go unattended, so the frontline can focus on exceptions and improvement.
🧭
STAKEHOLDER
HQ Ordering
Typically
Inventory checks and ordering easily consume several hours a day.
At RetailFlow,
Review the suggestion and complete ordering in one click.
Drastically cut ordering workload
Stockouts prevented too
🏬
STAKEHOLDER
Store staff
Typically
Arrival inspection tends to be matched visually against paper and PDFs.
At RetailFlow,
With auto-matching, you only need to check the discrepancies.
Lighter inspection load
Inventory accuracy improves too
🤝
STAKEHOLDER
Supplier
Typically
Orders tend to arise via phone, fax and manual entry.
At RetailFlow,
Accurate orders flow automatically to portals and EDIs.
Fewer ordering errors
Deliveries stabilize too
📊
STAKEHOLDER
Management / Operations
Typically
Person-dependent ordering work tends to become a black box.
At RetailFlow,
Ordering through arrival is visualized in one place and runs unattended.
A supply chain that never stops
Faster decision-making
05
DEMO
Three scenarios

Experience StockLink through a live, working demo.

FLAGSHIP
HQ Ordering
Order Management Console
  • · Auto-extract reorder SKUs from POS inventory and suggest order quantities
  • · Run RPA auto-ordering in one click
  • · Visualize order status in real time
Launch demo →
Store / Inspection
Arrival Matching
  • · Auto-match arrival data with orders
  • · Instantly highlight discrepancies (partial arrival / unmatched)
  • · Auto-reflect arrival results to inventory and master
Launch demo →
Operations / Admin
RPA Automation Tasks & Logs
  • · Centrally manage scheduled-batch timing and run status
  • · Integrate API-less external systems via screen operation
  • · Monitor unattended operation via execution logs and Teams/email notifications
Launch demo →