There’s a strategic reflex in software: see an incumbent, build a better version, take the market. In regulated agriculture that reflex is usually a trap, and it’s worth being explicit about why — because the discipline of what not to build is as much of the strategy as the product.

Some nodes are commodities. Don’t fight for a commodity.

Take the Certificate of Veterinary Inspection. It’s required to move animals across state lines, it’s increasingly electronic, and the systems that produce and submit it — the state-approved eCVI platforms — are genuinely good at that one job. They own the relationship with the state animal-health authorities. That relationship is the moat, and it’s theirs.

Trying to win there means out-executing a funded incumbent on a workflow they’ve owned for years, for a document that is fundamentally a commodity. You’d spend everything and, best case, become a slightly nicer version of a thing that already exists. That’s not a wedge. That’s a treadmill.

The same is true of the residue lab, the third-party certifier, the packer’s kill sheet. Each is the canonical source of its own record, and each has a reason to keep being that source.

The move: rent those rails, own the layer they can’t reach

Here’s the distinction that runs the whole strategy. There are things you should be the source of, things you should route to, and things you should facilitate.

  • Route/emit (rent the rails): the CVI, the lab result, the certificate. Don’t rebuild these. Produce the artifact, chain it into the record, help the vet submit it through the system the state already blessed, keep the clock on the submission window. You own the provenance and audit layer around the CVI — not the CVI.
  • Facilitate (the human authors): the nutrient-management plan a certified planner signs, every attestation a vet puts their name to. The software pre-fills and seals and stores; the professional authors and signs. That line is not a limitation — it’s what keeps document-correctness liability where it belongs and keeps the software out of practicing anyone’s profession.
  • Be the source (own the record): the things no incumbent actually produces — the treated-animal segregation record, the lot-eligibility determination, the continuous antibiotic-claim substantiation, the connective tissue that ties a movement to a claim to a treatment to an audit. That’s the open ground.

Notice what’s on the “own” list. It’s not any single document. It’s the graph — the connections between documents that every point tool structurally can’t see, because each point tool only holds its own node.

Why the connective layer is the durable one

A CVI platform emits a certificate and stops. It doesn’t bind that certificate to the movement record that also feeds the disease-traceability story, which also feeds the claim, which also runs alongside the VFD and the PQA. Each incumbent is excellent at its node and blind to the edges between nodes.

The edges are where the value that can’t be copied lives — because owning the edges requires holding all the nodes at once, and no point vendor does. It’s also where the switching cost lives: rip out a CVI tool and you lose a document generator; rip out the system that connects your CVIs, movements, treatments, claims, and audits into one chained record and you lose your operating history.

The honest posture this creates

This is also, not coincidentally, the honest posture. We don’t have to claim we replace the state’s system, because we don’t and shouldn’t. We don’t have to claim we are the lab or the certifier. We get to say, precisely: we emit the artifact and chain it; we help you submit it where it’s supposed to go; we own the record that ties everything together and makes the audit trivial.

Rent the rails. Facilitate the professionals. Own the record. The incumbents are safe in their nodes — and we’re building the one thing none of them can, which is everything in between.