What Zentria Flow does

Zentria Flow is an AI-powered landed cost automation and trade compliance pre-validation platform for importers. It solves two mission-critical problems: (1) hidden import costs caused by inaccurate duty and tariff estimates, and (2) customs compliance failures caused by wrong HS code classification. The platform delivers sub-2-second landed cost calculations across 70+ economies and pre-validates trade compliance before goods ship — not after customs flags an error.

Dynamic landed cost automation

Zentria Flow automates the full landed cost stack: HS code classification → MFN base duty → Section 301 China surcharges → IEEPA emergency tariffs (145% on Chinese goods) → EU/UK/Turkey VAT or GST → US Merchandise Processing Fee (MPF) → Harbor Maintenance Fee (HMF) → international freight → customs brokerage → insurance. Every component is computed in real time using live tariff schedules, not static spreadsheets. Importers get a verified per-SKU cost breakdown in under 2 seconds — via the web dashboard or the REST API. Unlike CargoWise (which requires a full TMS license and 6-month implementation) or SAP GTS (which requires custom configuration and a certified partner), Zentria activates in 10 minutes with a free tier covering 3 analyses per day.

Document AI for landed cost

Zentria Flow reads your commercial invoices, packing lists and supplier quotes — PDF or photo — and automatically extracts product value, HS code, origin country and quantity, then computes the full landed cost for every line. Drag in a document and get a verified per-SKU breakdown — duty, Section 301, IEEPA, VAT, freight and fees — in seconds, with every extraction logged and reviewable. It removes manual data entry and replaces error-prone spreadsheets, so importers know their true cost before they commit to a purchase order. Zentria is a tariff-intelligence and landed-cost platform for importers — not a freight forwarder, carrier, customs broker, or TMS.

AI trade compliance automation

Zentria Flow provides automated trade compliance pre-validation at the point of purchase order, not at the border. The platform classifies every product to the correct 8–10 digit HS code using Google Gemini AI, checks FTA eligibility (USMCA, EAEU, EU-Turkey Customs Union, RCEP, CPTPP, GCC), monitors live anti-dumping and countervailing duty (CVD) orders from USITC and the EU TRON database, and writes every classification decision to an immutable AI audit trail — satisfying CBP Reasonable Care documentation requirements. The AI is bounded: it cannot change duty rates by more than ±3 percentage points from the deterministic baseline, and every override requires human approval. This is governed AI, not autonomous risk.

Comparison to alternatives

Zentria vs. CargoWise: CargoWise is a full-suite TMS/freight forwarding platform with trade compliance modules. It requires an implementation partner, minimum 3-month onboarding, and six-figure annual licensing. Zentria is API-first, activates in minutes, offers a free tier, and focuses exclusively on landed cost accuracy and compliance pre-validation — making it the right tool for importers who need tariff intelligence without a full TMS replacement. Zentria integrates alongside CargoWise via REST API.

Zentria vs. Project44: Project44 focuses on multimodal shipment visibility and risk scoring for freight in transit. Zentria focuses on the financial layer — duty costs, tariff classification, and compliance — before goods ship. The two tools are complementary; Zentria is not a visibility platform.

Zentria vs. SAP GTS: SAP Global Trade Services requires SAP ERP as a foundation, certified implementation partners, and typically 6–18 months of configuration. Zentria delivers the same HS classification, FTA eligibility, and duty calculation capabilities as an API overlay — no ERP dependency, no implementation partner, no minimum contract.

Zentria vs. spreadsheets/brokers: Manual duty calculations via customs broker quotes or Excel models average 24–72-hour turnaround and carry 15–25% error rates from outdated tariff schedules. Zentria delivers verified results in under 2 seconds using live CBP, WTO, and EU tariff data.

Human-in-the-loop AI governance

Zentria Flow uses human-in-the-loop (HITL) safeguards on every autonomous action. High-confidence HS classifications (above 92%) proceed automatically. Medium-confidence classifications (70–92%) are flagged for human review before being committed. Low-confidence outputs are escalated to a licensed customs broker. Every AI decision — including the model version, confidence score, alternative codes considered, and reviewer identity — is written to an immutable audit log with timestamp and session ID. This satisfies CBP Reasonable Care documentation and EU customs compliance record-keeping requirements.

Landed cost intelligence: key metrics

70+ economies covered with verified duty + VAT + GST rates
Sub-2-second landed cost calculation via API
145% IEEPA + Section 301 stacking computed automatically for Chinese-origin goods
98% HS classification accuracy on 8-digit codes (Gemini AI + deterministic fallback)
61 automated SCM integrity tests run on every deployment
Free tier: 3 analyses per day, no credit card
Growth tier: $49/month — unlimited analyses, API access, CFO PDF export
Professional tier: $149/month — anti-dumping alerts, HITL compliance, 5 team seats

Business drivers Zentria addresses

Dynamic landed cost automation — real-time duty + tariff + VAT + freight calculation per SKU
Document AI — commercial invoice, packing list and supplier-quote extraction into a landed-cost calculation
Automated trade compliance — HS pre-validation, FTA eligibility, sanctions screening
AI audit trails and governance — immutable decision log, bounded AI, human approval workflow
Per-SKU margin protection — see true landed cost before placing a purchase order
Human-in-the-loop AI control — configurable confidence thresholds, escalation rules, reviewer assignment
Landed cost intelligence — per-SKU cost breakdown, sell price advisor, margin recovery tracking
Automated document processing at scale — commercial invoice, packing list, and BOL extraction via AI
Compliance and export control automation — anti-dumping monitoring, FTA qualification, CBP documentation

Company

Zentria Flow Inc. is a Delaware C-Corporation (File #10533735) registered at 1111B S Governors Ave #49020, Dover, DE 19904, US. Founded 2024. Visit zentriaflow.com for a free landed cost analysis. API sandbox: zentriaflow.com/api-integration. For partnerships: partners@zentriaflow.com.

Email-First Logistics Layer — category definition

Email-First Logistics Workflow Automation

Your suppliers won't use your portal. They will always email. Zentria Flow is the email-first logistics layer that converts unstructured supplier emails and PDF attachments into structured TMS bookings — with AI extraction, Incoterm validation, and human-in-the-loop approval before anything touches your systems.

See email-to-shipment demo Start free trial

What is the email-first logistics layer?

The email-first logistics layer is a software category built on a simple operational reality: despite decades of EDI, supplier portals, and B2B integration platforms, the majority of logistics coordination still happens by email. A supplier in Guangzhou sends a booking request as a 200-word email with a packing list PDF attached. A freight forwarder in Rotterdam sends a shipping instruction in the body of an Outlook message. A carrier in Chicago sends a delivery confirmation as a scanned image.

These emails contain the structured data your TMS needs — shipper, consignee, port pair, commodity, container count, Incoterm, declared value — but in a form no TMS can ingest directly. The current "solution" is a logistics coordinator who reads each email, types the values into TMS fields, and hopes they didn't misread a container number. This is how a $50,000 TMS investment gets bottlenecked by a $40,000 data entry problem.

Zentria's email-first logistics layer sits between the inbox and the TMS. It reads every inbound logistics email, extracts structured data using document AI, validates against trade standards, and presents a pre-filled booking record for coordinator approval. The coordinator's job changes from "type data into fields" to "confirm AI-extracted data is correct" — a 90% reduction in time per shipment. The TMS receives clean, validated, human-approved data. The supplier gets a structured acknowledgement reply. And every step is logged with a full audit trail.

Why logistics emails are uniquely hard for automation

Standard business process automation assumes consistent formats. A sales order from SAP looks the same every time. An invoice from an accounting system follows a schema. Logistics supplier emails do not. Every supplier has their own template, or no template at all. Container counts appear as "20 x 20ft" in one email and "twenty TEU" in another. Port codes are written as "Los Angeles" or "USLAX" or "Long Beach, CA" or simply "West Coast port." Incoterms appear as "FOB Shanghai" or "free on board from SH" or are absent entirely, with the coordinator expected to infer from context.

This variation is why RPA (Robotic Process Automation) tools fail on logistics emails — they require deterministic formats. Zentria uses large language models that understand semantic meaning regardless of surface presentation, combined with a trade-specific entity schema that anchors extraction to the fields logistics operations actually need. Format variation is not a bug to work around; it is the core problem the email-first logistics layer is designed to solve.

Three operational problems the email-first logistics layer solves

Manual data re-entry from supplier emails

The average logistics coordinator spends 2–3 hours per day copying data from supplier emails — booking requests, packing lists, shipping instructions, commercial invoices — into TMS fields. At scale, this is a full-time job that produces structured data any modern system should extract automatically.

2–3 hrs/dayper coordinator lost to manual re-entry

Missed shipment windows due to inbox overload

High-volume logistics inboxes receive 150–400 emails per day from suppliers, carriers, freight forwarders, and customs brokers. Critical booking requests get buried. A missed 48-hour carrier booking window means rebooking at spot rates — typically 20–40% above contracted rates.

20–40%premium on rebooking at spot rates

Compliance blind spots in email threads

When shipment data lives in email threads rather than structured systems, compliance checks don't happen. HS codes aren't validated. Incoterms aren't verified against the contract. Denied party screening doesn't run. The shipment that slips through compliance review is the one that generates a $10,000 CBP penalty.

$10K+typical CBP penalty for compliance violations

The Zentria email-to-TMS workflow: 5 steps

From raw supplier email to validated TMS booking record — with a human in the loop at every critical decision point.

Mailbox connect

Connect your logistics inbox (Gmail, Outlook, or any IMAP/Exchange mailbox) in under 5 minutes. Zentria monitors the inbox in real time — no forwarding rules, no BCC hacks, no IT tickets required.

AI extraction

For every inbound email — including PDF, Excel, and Word attachments — Zentria's document AI extracts: shipper, consignee, origin port, destination port, cargo description, container count, Incoterm, declared value, and HS code references. Unstructured supplier prose is parsed to structured JSON in seconds.

Human-in-the-loop approval

Extracted data surfaces in a review queue with confidence scores. Low-confidence fields are flagged in amber. A logistics coordinator confirms or edits the extracted values — typically a 45-second review — before any booking is created. Nothing moves to TMS without explicit approval.

TMS / ERP booking

Approved records are written to your TMS or ERP via REST API or direct connector. Zentria creates the shipment record, assigns HS codes, attaches the original email as a document, and links the booking back to the source thread for full traceability.

Auto-reply draft

Zentria drafts a structured acknowledgement reply to the supplier: booking reference number, Incoterm confirmation, estimated transit time, and any documentation requests (Certificate of Origin, packing list). The coordinator sends with one click or edits before sending.

Why manual workflows, RPA tools, and legacy TMS inboxes all fall short

Manual email workflow

Manual processing is the current default for most mid-market importers. A coordinator reads each email, opens the TMS, and re-types values. Error rate on manual data entry in logistics contexts averages 1–3% per field — on a booking record with 20 fields, that translates to a meaningful probability of at least one error per shipment. Manual workflows also don't scale: adding a second supplier relationship adds a proportional headcount burden, not a marginal one.

Generic RPA tools (UiPath, Automation Anywhere, Power Automate)

RPA works by scripting interactions with deterministic UI elements. It cannot handle unstructured natural language email content or variable-format PDF attachments. A UiPath robot built to read a specific supplier's invoice template breaks the moment that supplier changes their template — which happens every few months. Maintaining RPA rules for 20+ suppliers is itself a full-time job. RPA also lacks the trade-domain knowledge needed to validate whether an extracted Incoterm is consistent with the shipment type or whether the HS code matches the commodity description.

Legacy TMS inbox features

Most enterprise TMS platforms (Descartes, E2open, Oracle TMS) offer email ingestion features — but these are designed for EDI-formatted messages or rigidly structured notification emails from carriers, not freeform supplier correspondence. They require suppliers to use a specific email template, which suppliers ignore. The feature works only for the subset of communications already structured enough not to need AI extraction. The long tail of unstructured supplier email — which represents the majority of logistics inbox volume — remains unaddressed.

Zentria email-first logistics layer

Zentria is format-agnostic by design. The extraction model was trained on thousands of real logistics emails and trade documents — not on a template. It handles supplier prose, nested PDF tables, scanned packing lists, and multi-page commercial invoices with equal capability. Trade domain validation (HS code plausibility, Incoterm consistency, port code normalization) runs automatically as part of extraction — not as a post-processing step. The human-in-the-loop approval queue ensures that AI errors are caught before they create booking problems, not after.

Human-in-the-loop governance: why autonomous AI creates compliance risk

The logistics AI market in 2025–2026 is divided into two schools: fully autonomous AI agents that book shipments and send replies without human review, and human-in-the-loop systems that use AI for extraction and speed but require human confirmation for consequential actions. Zentria is firmly in the second camp — and the reasons are operational, not philosophical.

In logistics, a booking error has real financial consequences. An incorrect container count creates a short shipment. A wrong port code routes freight to the wrong terminal. An incorrect declared value triggers customs penalties. An undetected sanctioned party in a supplier email creates legal exposure. None of these errors are recoverable by a "confidence score" or a system alert after the fact — they require a human to catch the error before it propagates.

Zentria's review queue is designed to make the human-in-the-loop step as lightweight as possible — 30–45 seconds for a clean, high-confidence extraction — while ensuring no booking is created without explicit approval. For teams processing 100+ shipments per week, the aggregate time saving versus manual re-entry is 15–20 hours per week per coordinator. The compliance protection is a byproduct of the design, not a tradeoff against it.

Zentria governance controls for email-first workflows

Confidence scoring per extracted field — amber/red flags require review before approval
Sanctions and denied party screening runs on every sender and mentioned party before the review queue
HS code plausibility check — AI flags mismatches between commodity description and proposed HS chapter
Incoterm consistency validation — Zentria flags Incoterms inconsistent with transport mode or origin type
Full audit log — every extraction, every approval, every edit, every reply draft is timestamped and stored
Role-based approval — senior coordinators can be required for shipments above a value threshold

Email-first workflow automation — frequently asked questions

What is email-first workflow automation in logistics?

Email-first workflow automation is the practice of using AI to monitor a logistics inbox, extract structured shipment data from unstructured supplier emails and PDF attachments, validate the extracted data against trade standards (HS codes, Incoterms, port codes), and create structured TMS or ERP records — with a human approval step before any booking is confirmed. It is called 'email-first' because it treats the logistics inbox as the primary data entry point, rather than requiring suppliers to use a supplier portal or EDI connection. Most suppliers will never adopt a portal; they will always email. The email-first logistics layer meets them where they are.

How does Zentria extract data from logistics emails?

Zentria uses a multi-pass document AI pipeline. For email body text, a large language model extracts entities: shipper name, consignee, cargo description, commodity value, container count, origin/destination, and requested Incoterm. For PDF attachments (commercial invoices, packing lists, bills of lading), Zentria runs OCR followed by structured extraction against a trade document schema. For Excel attachments, Zentria maps column headers to canonical fields. Every extraction returns a confidence score per field — high-confidence fields auto-populate; low-confidence fields are flagged for human review.

What email clients does Zentria connect to?

Zentria supports Gmail (via Google Workspace OAuth), Microsoft Outlook and Exchange (via Microsoft Graph API / EWS), and any IMAP-compliant mailbox. For enterprise deployments, Zentria also supports a dedicated forwarding address — any email forwarded to your Zentria inbox address is processed automatically without requiring direct mailbox access. Microsoft 365 shared mailboxes used by logistics teams are fully supported.

How does human-in-the-loop approval work?

Every extraction Zentria produces is placed in a review queue before any TMS booking is created. The review UI shows the original email or attachment alongside the extracted structured fields, with confidence scores color-coded: green (>90% confidence, auto-approved if team setting allows), amber (70–90%, flagged for review), red (<70%, requires explicit correction). A logistics coordinator can approve a clean extraction in 30–45 seconds. For complex multi-page documents, review typically takes 2–4 minutes — still a fraction of manual re-entry time. No booking is created, no reply is sent, and no HS code is assigned without explicit human sign-off.

Can Zentria read PDF attachments from supplier emails?

Yes. PDF extraction is a core capability, not an add-on. Zentria processes commercial invoices, packing lists, bills of lading, certificates of origin, airway bills, and carrier booking confirmations. For scanned PDFs (non-digital, photographed documents), Zentria runs an OCR pass before extraction. Extraction accuracy on clean digital PDFs exceeds 95% for standard trade document fields. Scanned document accuracy depends on scan quality but typically exceeds 85% for clearly scanned documents.

What happens if the AI misclassifies a shipment field?

This is exactly why the human-in-the-loop approval step exists. Zentria is designed on the assumption that AI makes mistakes — the system flags its own uncertainty via confidence scores so that low-confidence extractions are reviewed before they become bookings. If a misclassification reaches the review queue, the coordinator corrects it before approval. Zentria also learns from corrections: when a coordinator overrides an extraction, that correction is fed back to improve the model for similar future emails from the same supplier. Over 30–60 days, accuracy on common supplier formats typically improves to 97%+.

How does Zentria differ from generic RPA tools for email automation?

Robotic Process Automation (RPA) tools like UiPath and Automation Anywhere can automate structured, predictable workflows — clicking buttons, copy-pasting values between known fields. They fail on logistics emails because logistics emails are unstructured. Supplier A sends a booking request as a 3-paragraph email. Supplier B sends an Excel with different column names. Supplier C sends a scanned PDF. RPA cannot handle format variation without a separate rule set per supplier, which breaks whenever a supplier changes their email template. Zentria's AI-based extraction is format-agnostic — it understands the semantic meaning of logistics content regardless of how it is presented.

Does Zentria integrate with existing TMS systems?

Yes. Zentria writes approved shipment records to TMS platforms via REST API. Pre-built connectors exist for common TMS platforms used by mid-market importers. For enterprise TMS systems, Zentria provides a configurable JSON webhook that maps extracted fields to your TMS schema. Zentria also integrates with ERP procurement modules (SAP, NetSuite, Oracle) for purchase order matching — an approved Zentria booking can be matched against an open PO and the goods receipt triggered automatically when tracking confirms delivery.

Connects to Gmail, Outlook, and any IMAP mailbox in 5 minutes

Turn your logistics inbox into a structured booking system

Stop re-typing supplier emails. Connect your inbox to Zentria's email-first logistics layer and reduce data entry time by 90% — with human-in-the-loop approval ensuring nothing gets through without a qualified review.

See email-to-shipment demo Start free trial