Multi-tenant WMS · Ecommerce intelligence

The warehouse ledger and the demand signal, in one system.

SysTIAM runs the full flow of goods through your warehouse — receive, put away, allocate, pick, pack, ship, return — and holds sell-through from your sales channels in the same multi-tenant database. Warehouse platforms usually receive demand as an opaque order feed. This one can see why the orders arrived.

Built for multi-client third-party logistics providers, and for the brands they serve.

7: Sales-channel connectors, from Shopify to SAP Commerce Cloud
850 · 856 · 945 · 810: EDI documents, with 997 acknowledgements returned automatically
SHA-256: Hash-chained audit ledger, re-verifiable through the API on demand
EN · 繁中: Both interface languages held at parity by a build-breaking test

One ledger, from the dock door to the doorstep

Receive Put away Allocate Pick Pack Ship Return

The everyday cycle is the same one your floor already works to. What differs is underneath: every stock change — a receipt, a transfer, a re-slot, an assembly, a pick — passes through a single movement chokepoint. That makes the inventory record forensic rather than approximate, and it is why a count, a recall or an invoice can be traced back to the movement that caused it.

Two jobs, one platform

The warehouse and the storefront are usually two systems that exchange files. Holding both means a decision on one side can be made with the other side's facts.

For multi-client 3PLs

Segregation enforced, not promised

Client ownership is carried through allocation, picking, holds, receiving, transfers and assembly. One client's units cannot quietly fill another client's order.

Billing as a by-product of operations

Work performed becomes a billable event, priced against a rate card and rolled into an invoice PDF — storage, handling, value-added services and freight — with client wallets, plans and multi-currency settlement.

Freight and the yard, not only parcels

LTL rate cards by freight class with weight breaks and a minimum floor, flat FTL lanes, multi-stop load planning, dock appointments, and trailer dwell and detention clocks.

Proof before anyone commits

A reconciliation engine compares your incumbent system's records against ours, day after day, across every kind of record — so a cutover is a decision made on evidence.

For brands and merchants

Reorder driven by sell-through

Reorder points and demand forecasts read what is actually selling on the channel, not warehouse pick history alone, and turn it into a purchase recommendation.

Stockouts ranked by revenue at risk

The same SKU is tracked as it runs out on the listing and as it runs out in the bin, and surfaces as one alert ordered by what it is costing you.

Listings and stock kept in step

Create and manage listings across channels, including multi-variant, with two-way inventory sync that detects channel-side edits and heals them.

Post-purchase that stays yours

Branded tracking pages, email and SMS notifications, a self-serve returns portal, and exception alerts on shipments that have stopped moving.

What runs the floor

Every capability below has a working screen behind it, named so you can go and look at it.

Inventory accuracy and visibility

Balances by location, lot and client, serial units, holds, FEFO and recall on lots, GS1 product identifiers, and RFID reconciliation against the ledger.

/inventory/balances

Inbound and quality

Purchase orders and receipts, directed putaway, quality inspection, cross-docking, dock scheduling, and yard moves from gate to door.

/receiving

Picking

Waves with rule-based auto-release and waveless streaming, allocation strategies, pick-path and zone-parallel routing, an RF worklist, batch picking, and pick- and put-to-light.

/worklist

Packing and shipping

Cartonization, pack verification with photo evidence, rate shopping, manifests, bills of lading, SSCC and GS1-128 labels, and return labels.

/pack-station

Freight and load planning

Freight class on the SKU, LTL and FTL rate cards, a quote engine, multi-stop trailer planning, dispatch under a PRO number, and a bill of lading per load.

/loads

Slotting and labour

ABC and affinity slotting that executes as a re-slot task on the floor, engineered labour standards, interleaved next-task assignment, and expected-versus-actual variance.

/slotting

Kitting and value-added services

Assembly work orders over a bill of materials: availability-gated release, work that hits the same worklist as a pick, FEFO consumption, and the charge raised exactly once.

/assembly

Multi-site networks

Transfer orders with a conserved in-transit pool, internal ASN receiving at the destination, per-SKU availability across the network, and order sourcing between sites.

/transfers

Control tower and SLA

Orders at risk, blocked work, open alerts and platform health in one worst-first cockpit, with deep infrastructure checks and monitors that raise an alert when a target is breached.

/control-tower

Built to be audited

Institutional buyers do not accept assurances; they ask what evidence you can hand over. These are controls with code behind them and tests that hold them in place.

An append-only ledger

Entries are SHA-256 chained to the one before them. The model refuses any update or delete outright, and the chain can be re-verified through the API at any time — so integrity is demonstrable on demand rather than asserted in a document.

One authorisation chokepoint

Every protected endpoint passes through a single guard carrying a role set and a scope list. A test asserts that seventeen write endpoints reject a read-only user, which makes the guarantee checkable rather than architectural folklore.

A second person, by identity

Four actions — invoice adjustments, cycle-count completion, client cutover and elevated-role grants — require a different human being. An approval by the requester is refused before roles are even considered, and API keys are refused outright.

Isolation that covers the jobs too

Every domain model is tenant-scoped, and the isolation tests cover background tasks as well as requests — which is where cross-tenant bugs actually hide.

Secrets and personal data

Credentials encrypted at rest with key rotation. Buyer data redacted at the integration boundary, purged on a retention schedule, and erasable on request behind a double confirmation.

Monitors with a track record

The SLA monitors over the asynchronous backbone caught a real gap in a running system: an event outbox with no scheduled consumer, silently backed up. A monitor that has caught something is better evidence than one that has only ever been green.

What this is not

SysTIAM holds no SOC 2 or ISO attestation, and nothing here should be read as implying one. What exists is the engineering half of that work: the controls, the evidence they generate, and an internal readiness assessment that is candid about the gaps that remain.

It has to talk to what you already run

Connections are built on a shared connector base with retries, circuit breaking, a dead-letter queue with replay, and per-call telemetry — so an integration that degrades is visible before a client notices.

Sales channels

Shopify · WooCommerce · BigCommerce · Magento (Adobe Commerce) · Salesforce Commerce Cloud · SAP Commerce Cloud · HKTVmall

3PL and WMS systems

Extensiv (3PL Warehouse Manager) · ShipHero · ShipBob · ShipMonk

Carriers and tracking

FedEx · DHL · UPS · AfterShip universal tracking

EDI

850 purchase order · 856 advance ship notice · 945 shipping advice · 810 invoice · 997 acknowledgement, over SFTP or a filesystem inbox

Coverage is per capability, not uniform, and we would rather say so here than in a call. Connectors run against a deterministic simulator until live credentials are supplied. Of the carriers, FedEx label purchase is validated against the production API; the remaining calls await carrier account verification. Salesforce Commerce Cloud and SAP Commerce writes are built and simulator-tested, pending a client sandbox. EDI 940 and AS2 are not implemented.

Switching systems is the risk, so it is the part that is engineered

Most warehouse migrations are a weekend, a spreadsheet and a held breath. This one is a sequence with a gate at each step, and a rollback that has been rehearsed rather than described.

1
Keep your integration

If your code already calls the Extensiv API, point the base URL at the compatibility façade and swap the auth header. It accepts those payloads and answers in the same shapes. Move to the native API when it suits you — not on cutover day.
2
Mirror

Both systems hold the same data while the incumbent stays authoritative. Nothing changes for the operation on the floor, and nothing is at stake yet.
3
Parallel run

The reconciliation engine compares every kind of record. The flip stays locked until each of them matches at 99% or better for three consecutive runs — a gate the software enforces, not a judgement call.
4
A governed flip, then hypercare

The transition is a typed confirmation against a readiness gate, written to the audit ledger — not a quiet database update. Afterwards billing runs in dual-run against the incumbent's charges, to the cent, while you watch.

See it on real data

The demonstration environment is seeded by the same services that run the warehouse. Every number on screen is produced by the real ledger, not painted onto a screenshot — which also means you can break it, and it will tell you honestly.