Harborlane Freight

For freight forwarding and customs brokerage operations

Shipping documents, read once and checked by a person

Harborlane Freight sets up document workflows for operations teams handling commercial invoices, bills of lading and packing lists. Documents are filed by shipment in SharePoint, key fields are extracted with Azure AI Document Intelligence, and a member of your team reviews every result before it’s used anywhere else.

  1. Shipment-document organisationEvery document in the right shipment folder
  2. Field extractionAgreed fields read from each document type
  3. Human review and exception routingNothing moves on until someone has checked it

Azure AI Document Intelligence · Power Automate · SharePoint

Tell us about your document mix

A technology services business. We don’t provide freight forwarding or customs brokerage, and extraction doesn’t make a declaration compliant.

From inbox to downstream system

How a document moves through the workflow once it’s set up. Step four is the point of the design: extracted values are suggestions until a person confirms them.

  1. ReceivedArrives by email or a customer portal, as it does today.
  2. FiledSaved to the shipment’s folder in SharePoint, with a consistent name and document type.
  3. ExtractedAgreed fields are read, each with a confidence score.
  4. Checked by a personThe reviewer sees the document beside the values, then confirms or corrects them.Required before useLow confidence, missing fields or mismatches go to a named exceptions owner.
  5. ApprovedValidated values are recorded with who checked them and when.
  6. Handed onPassed to your forwarding or customs software by the route your team chooses.

Illustrative. Steps and thresholds are designed with your team, and the hand-off in step six depends on what your other systems accept.

Services

Three pieces of work that build on each other. Organisation comes first, because extraction and review both depend on knowing which shipment a document belongs to.

Shipment-document organisation

SharePoint · Power Automate

A SharePoint structure where each shipment or job has one home, and documents arriving in shared mailboxes are filed there automatically, under a consistent name, instead of living in individual inboxes.

Deliverables

  • Folder and metadata design: job reference, mode, customer code, document type
  • Automatic filing of attachments from agreed shared mailboxes
  • A naming convention and a short guide for the team
  • Access by team or branch, and retention set to your policy

Field extraction from shipping documents

Azure AI Document Intelligence · Power Automate

Prebuilt models where they fit, such as invoices, and custom models trained on your own document layouts where they don’t. Each extracted value carries a confidence score, which decides how the review step treats it.

Example fields only. The list is agreed per document type and tested on your samples.
DocumentTypical fields
Commercial invoiceInvoice number and date, seller, buyer, currency, Incoterms, line descriptions, quantities, values, total
Bill of ladingB/L number, shipper, consignee, notify party, vessel and voyage, ports of loading and discharge, container and seal numbers
Packing listPackage count and type, gross and net weights, dimensions, marks and numbers

Deliverables

  • A field list and confidence threshold for each document type
  • Models set up and tested on a sample of your documents, with the results reported as measured
  • Extracted values stored alongside the document they came from

Accuracy varies with layout, scan quality and language. We report what we measure on your documents rather than promising a figure, and the review step is designed for the cases extraction gets wrong.

Human review and exception routing

Power Automate · SharePoint · Power Apps or lists

A review queue where a team member sees the document next to the extracted values, confirms or corrects them, and approves. Anything unusual is routed to a named person rather than left in a shared inbox.

Deliverables

  • A review screen showing document and values side by side
  • Routing rules: low confidence, missing required fields, values that disagree between documents
  • An approval record: who checked, what changed, when
  • A rule that only approved values are passed on

Two exceptions, followed through

How the review step handles the cases that matter. Hypothetical examples, not client documents.

Illustrative

Invoice total doesn’t match its lines

  1. Extraction reads a total that differs from the sum of the line values.
  2. The mismatch rule routes the document to the exceptions owner for that customer.
  3. They check the original, correct the value or query the supplier, and record why.
Illustrative

A container number read with low confidence

  1. The scan is faint and the confidence score falls below the agreed threshold.
  2. The reviewer sees the value highlighted beside the image of the document.
  3. They confirm the number against the original before approving.

Extraction speeds up reading. It doesn’t replace the judgement of the people responsible for each shipment.

Your licensed brokers and operations staff stay responsible for every declaration and booking. The workflow’s job is to make sure a person has checked the data before it’s used.

Azure AI Document Intelligence is billed through an Azure subscription, usually by pages processed and model type. We estimate costs from your volumes before anything is built.

Shipment Document Workflow Checklist

A printable checklist covering intake, filing, extraction set-up, human review, exceptions and hand-off. Use it to describe how documents move today, or to agree how they should move before any automation is added.

Open the checklist

About Harborlane Freight

Harborlane Freight works on the document side of freight operations: where shipping documents are kept, how their contents are read, and how people check them before they’re relied on. “Freight” in the name describes who we work with. We don’t move freight or act as a customs broker.

  • Not a freight forwarder, carrier or customs broker
  • No promise of extraction accuracy or customs compliance
  • No customs, tariff-classification or trade-compliance advice
  • Independent; not affiliated with or endorsed by Microsoft

Tell us about your document mix

Which document types you handle most, roughly how many a week, and where they arrive today. A general description is plenty.

Please don’t send confidential shipment documents, customer details, identity documents or login credentials. The form doesn’t accept attachments.

Send an enquiry

Fields marked are required.

Describe document types and volumes in general terms. No customer or shipment details. Up to 3,000 characters.

0 / 3,000

We use your details only to reply. See the privacy policy.