Produce traceability a two-person packhouse can run

PTI compliance for small growers with GS1 company prefix registration bundled in, guided first label runs, mobile scanning, receiving, and a month-end QuickBooks export. A grower meets a retail buyer's mandate without hiring a compliance manager.

Client
AgTag USA
Period
2025–26
Our role
Product design, full-stack build, AI features
Built with
PHP, MySQL, jQuery

The problem

When a retail buyer hands a grower a Produce Traceability Initiative mandate, the grower's options have historically been to buy enterprise traceability software built for operations fifty times their size, or to lose the account. The compliance requirements do not scale down, but the available tooling assumed a compliance department that a two-person packhouse does not have.

What we built

A traceability platform sized for small growers, where the hard parts are handled rather than delegated back to the customer.

  • GS1 company prefix registration included in every plan, not sold as an add-on or left as homework. This is the single most common place a small grower stalls.
  • Case label generation that produces compliant PTI labels with the right identifiers, lot codes, and dates.
  • Mobile scanning and receiving for the packhouse floor, on the phone that is already in someone's pocket.
  • A QuickBooks export as a flat spreadsheet that finance can actually import at month end, with PTI and non-PTI order logic handled correctly.

How it works

Guided onboarding instead of a manual

A new account answers five questions - what they grow, how they pack, who they ship to, how often, and whether they already have a GS1 prefix - and the account configures itself from the answers. Then it walks the grower through their first real label run conversationally. Nobody reads a PDF manual to get started.

Photograph it instead of typing it

Paper still runs a lot of packhouses. Capture a packing slip on a phone, and a vision model extracts the fields with a confidence score per value. The extraction always lands on a confirmation screen, and the original image is retained as part of the audit trail.

A rule we do not bend

No AI extraction reaches a compliance record without a human confirming it. In traceability, a confidently wrong lot code is worse than a blank field, because the whole point of the system is that the record can be trusted during a recall.

Framed by outcome, not by technology

None of these features are marketed as "AI" to this audience. They are described by what they remove: less typing, a faster first label, fewer support calls in week one.

What it changed

A grower who was facing a compliance ultimatum can now satisfy a retail mandate with the staff they already have, and keep the account.

5Onboarding questions before setup completes
1Confirmation screen before any AI data is saved
14Columns in the finance-ready export
0Compliance managers required
PHPMySQLjQueryGS1 & PTIVision APIXLSX export
Get in touch

Tell us what the ground looks like.

A few sentences about the program, the data you already have, and the deadline you are working against is enough to start. You will hear back from the person who would do the work.

What happens next
First

A real reply, not a ticket number

Usually within one business day, from the person who would build it.

Then

A short call about the data

What exists, where it lives, and who has to use the result.

After that

Scope and a fixed number

Written scope with the cuts called out, so the budget decision is yours.