Linked operating records from industries your models haven't seen

We license de-identified email, system-of-record and accounting history from US freight brokerages, manufacturers and distributors, specialty contractors, commercial insurance agencies and customs brokers. Records are delivered as linked workflow chains with outcomes.

Aerial view of a shipping container yard

Where we are: we're assembling our first sector corpora now, starting with US freight brokerage. Tell us what you need and we'll share what's in our pipeline under NDA.

What a delivery includes

ComponentWhat it contains
Workflow chainsMessages, documents and system records joined by shared IDs (load, quote, job, policy or entry number), with outcomes such as won or lost, paid or disputed. Delivered as JSONL.
AttachmentsRate confirmations, quotes, ACORD forms, change orders and other PDFs, with de-identified extracted text.
Datasheet and schemaCounts, date ranges, field definitions, source systems and known gaps.
De-identification reportMethods used, hand-checked sample results, and motivated-intruder findings.
Rights summarySeller warranties, consent records where required, and what was excluded and why.

Licensing options

Non-exclusive

The default. Training and evaluation use for an agreed term.

Time-limited exclusive

Six to twelve months of exclusivity on a corpus, priced at a premium.

Evaluation only

Held-out sets for benchmarks and agent evaluation, never used for training.

What we require of buyers

  • Use limited to training and evaluating AI models
  • No attempt to re-identify any person or company
  • No resale or onward transfer
  • Deletion when the license ends
  • A completed security questionnaire
  • We don't license to entities in countries of concern under federal bulk-data rules, and foreign buyers agree to matching onward-transfer limits

Request a data brief

Tell us which sectors and workflows you need. We'll reply with what's in our pipeline, under NDA.

Data brief request

Sectors of interest