Integration

Shelfbot integrates with your existing systems via secure API so products, inventory and orders stay in sync — and bins are presented accurately for picking and replenishment.

Key features:

• WMS/ERP agnostic
• JSON payloads
• Designed for operational clarity

---

What data flows between systems

Shelfbot needs three things to run smoothly: product master data, inventory, and orders.

Products (SKU master)

Shelfbot imports SKU master data so items can be searched, stored, and picked reliably (SKU, description, barcodes; dimensions/weights preferred).

Inventory

Inventory movements (replenishment, adjustments, cycle counts) are reflected between systems so Shelfbot knows what is available to pick.

Orders

Orders and line items flow from your WMS/ERP into Shelfbot, which then requests bins and confirms pick completion back to your system.

---

Minimum integration requirements

Unique identifiers

• Unique SKU (including variants)
• Unique order ID
• Unique order line ID

API capability

• A documented API with secure access (cloud-based or accessible from your network)
• Ability to create/update/delete products (or at minimum create/update)
• Ability to create/update inventory transactions
• Ability to create/update orders and receive order status updates (or accept status polling)

Operational discipline

• Operators scan every pick and replenishment (to prevent stock drift)
• Stock on hand in the external system is accurate before enabling live picking

Practical note: Inventory accuracy depends on scanning discipline. If items are removed or added without scanning, inventory will drift between systems.

---

Integration approaches

If you haven’t built an API yet, start with polling and move to webhooks/events as you mature.

1) Simple polling (fastest start)

Shelfbot periodically pulls orders and product/inventory updates from your system. Good when webhooks aren’t available yet.

2) Webhooks / event-driven (recommended)

Your system pushes changes to Shelfbot as they happen (new order, inventory update, SKU update). Lower latency and less API load.

3) Middleware / iPaaS

Use an integration layer to map fields, transform payloads, and handle retries/monitoring. Helpful for multi-system environments.

---

Shelfbot API

The Shelfbot API documentation is shared with clients through a secure link during the integration process.

---

Integration process

Discovery

• Confirm your systems (WMS/ERP/e-comm) and integration constraints
• Agree on the “source of truth” for product, inventory, and orders
• Identify barcode standards and any edge cases (bundles, kitting, substitutions)

Mapping

• Map your SKU, order and inventory fields to Shelfbot’s model
• Define status events and idempotency keys
• Define error handling (retries, dead-letter queue, human review)

Sandbox & testing

• Test with sample orders and known inventory
• Validate pick confirmation, partial picks, and replenishment flows
• Run a short UAT with operators using real scanners and stations

Go-live

• Cutover plan and monitoring
• Operator training (pick, replenish, search, exceptions)
• Post-go-live tuning (slotting, batching, FIFO strategy as needed)

---

Start fast with a data dive

Before you write a full integration, we can size the system using SKU master data + outbound orders. That gives you confidence on bin count, aisle count, and robot/station sizing.

Quick checklist

• Do you have unique SKUs and barcodes for pick confirmation?
• Can your system export 30+ days of order history (ideally 12 months) for sizing?
• Do you have a stable “source of truth” for inventory on hand?
• Are there any DG / restricted items you plan to store (we’ll confirm what’s permitted)?
• Do you need one-way (orders → Shelfbot) or two-way (status + inventory sync) integration?

Download data dive specification (PDF)

---

Ready to map your system?

We’ll keep it practical: agree on the source of truth, map the fields, test with real scanners, then go live.

Talk to an engineer • Product overview • Safety & compliance