Cross-tenant traceability
From a banned active to the buyer who received it — in one traversal.
A regulator withdraws approval for an active substance on a Tuesday. Which of your European buyers is holding produce exposed to it? That question crosses a tenant boundary, jumps between two schemas, and has no foreign key to lean on. It is the reason SoftFarm keeps a graph.
The problem
Two tenants, no foreign key between them
The producer owns one chain: Crop → Application → Product → Substance, plus residue tests and dispatches. The exporter owns another: GoodsReceipt → PackRun → Pallet and PackRun → SalesOrder → Buyer.
One consignment’s real history runs through both. Relationally that is an unbounded recursive join across two schemas with a cross-tenant hop that referential integrity does not cover — the exact workload a property graph exists for.
So SoftFarm keeps a graph for those queries and only those queries. It is not a second database. It is a read-optimised projection of the relational truth.
One question, one answer
Seven hops across two separate businesses, returned as one answer — in the app, as a diagram you can tap through.
The model
The spine of the graph
Below is the path that the exposure query walks. The full projection carries more — targets, residue tests, pallets, storage zones, customers — but this is the backbone, and the one edge that crosses tenancy.
The chain is recorded as the work happens, so answering a question about it is a lookup rather than an investigation.
What makes it dependable
Answers you can act on
Traceability is only worth having if you trust it under pressure — on the day a buyer calls, not in a demo.
Recorded, not reconstructed
Every link is written at the moment the work happens — the spray, the delivery, the pack run, the pallet. Nothing is inferred later from paperwork.
Across the boundary
The farm and the packhouse are separate businesses with separate accounts. What crosses between them is the delivery, and it carries its history with it.
Both directions
Start from a chemical and find the buyers, or start from a buyer complaint and find the field. The same chain reads either way.
Honest about gaps
Produce from a supplier who is not on the platform carries less history, and the system says so rather than presenting a confident but incomplete picture.
In the product
Rendered in-app, not in a database console
The exporter portal’s Supply Graph and Trace & Recall pages draw the chain as a diagram you can tap through. No separate tool, no export to a spreadsheet, no waiting on anyone.
Compliance exposure
Every finished lot whose upstream applications used a substance the EU has not approved — with the buyers and orders that received it.
Lineage explorer
Enter any code — lot, pallet, order, receipt — and walk the chain forwards to buyers or backwards to fields.
Reverse recall
Select by lot, producer, crop or date range and resolve the full affected set before you commit to a response.
Supplier risk
Producers ranked by residue failures, non-approved applications and reject attribution against their dispatches.
Buyer concentration
How much of your volume depends on how few buyers — a commercial risk the same graph happens to answer.
Producer centrality
Which suppliers are load-bearing across many lots, and would hurt most if one of them failed an inspection.
Co-mingled lots
Which farms ended up in the same finished lot — the reason one bad supplier can contaminate a clean one commercially.
Co-location risk
Lots that shared a storage zone, for the contamination questions cold rooms create.
Mass-balance leaks
Points where the quantities stop reconciling — usually a missing record rather than missing produce, and worth finding either way.
The full model, the query catalogue, the verification steps.
How to trace a carton back to the field, how to find every buyer affected by one supplier, and what to do when a consignment is queried — step by step in the user guide.