PRODUCT · ORDERS · EXCEPTION HANDLING

When something goes wrong, the system already knows who needs to fix it.

Trade orders break in dozens of ways — vessel delays, QC failures, missing certificates, FX shocks, manufacturer no-shows, customs holds. Most operators learn about these by accident, three days late, when a client asks an awkward question. TradeOS detects exceptions the moment they happen, classifies them by impact, routes them to the right person, and tracks resolution end-to-end. The audit trail makes “what went wrong” answerable in seconds.

Book a demo See pricing →

TradeOS

HomeTasksExceptionsOrdersProductionShipmentsFinance

Exception inbox ·8 active·3 high severity·14 resolved this week

AllMineHighAwaiting meResolved

E-204

Sapphire QC fail

Lot PL-2026-0142 · 2 major defects

$231K4h ago

E-202

Missing Certificate of Origin

RD-2026-077 · blocks customs

$560K8h ago

E-198

Athena margin compression

19.8% projected vs. 22.5% target

$1.25M1d ago

E-203

Vessel delay COSCO Pacific

+2 days · 6 orders affected

$14.4K12h ago

E-199

Manager approval stuck

RD-2026-091 · Lotus Rubber

$340K1d ago

E-197

SafeHands invoice past 90d

3 reminders sent

$36K2d ago

E-200

Sapphire EUR.1 outstanding

Auto-reminder sent

0 hard3d ago

E-201

CE cert expiring

Crescent · Protect-M Nitrile

0 hard47d

E-204Sapphire QC failHIGH

Detected May 5, 09:14 · Owner: Liam Okafor · Status: Awaiting decision

SourceProduction lot PL-2026-0142

InspectionINS-0249 · In-process AQL 1.5

Sample size200 units

Defects8 minor · 2 major · 0 critical

AQL thresholdFAIL · max 5 major (over by 1)

ManufacturerSapphire Industries 🇹🇭

OrderRD-2026-077 · Brevin Health EU

Lot value$231,2003.2M units × $0.072

Downstream1 shipment · 1 invoice · 1 customs declaration

Event timeline

09:14QC inspection submitted by External AQL Co.4h ago

09:14Auto-classified HIGH severity4h ago

09:14Assigned to Liam Okafor (Operations)4h ago

09:18Push notification sent to Liam4h ago

09:34Liam acknowledged3h 40m

RD-2026-077· Brevin Health EUIN PRODUCTION

Expected delivery: Jul 21 → Jul 26 if reworked, Jul 21 → Jul 28 if re-allocated.

Margin impact: 21.2% → 19.8% (rework) or 21.2% → 18.4% (re-allocate).

Resolution options

Each option shows downstream impact

RECOMMENDED

Impact+3 days · $4,200 rework cost · margin 19.8%

Recommended by pattern history (3 of 4 similar cases reworked successfully).

Impact+7 days · $8,400 re-allocation cost · margin 18.4%

Trade-off: longer delay but higher confidence.

Impactindeterminate · pending response · clock keeps running

Use when: pattern of recurring quality issues warrants direct conversation.

Whichever option is chosen writes to the audit trail with reason and impact assessment.

THE PROBLEM

Most exceptions surface in inboxes, three days late.

A QC inspector emails a PDF report. A vessel delay shows up in the carrier portal but no one’s logged in. A manufacturer flags a material shortage in WhatsApp. The customs broker mentions a missing document on a phone call. None of these become events in any system. The operator finds out when the client asks why their order shifted, or when an invoice gets disputed, or when the monthly margin report shows compression no one explained. By the time the exception is visible, the resolution window is closed.

SIX CATEGORIES OF EXCEPTIONS

Trade breaks in patterns. The platform names them.

Every trade exception fits one of six categories — production, shipment, document, financial, compliance, or counterparty. Each category has a defined detection method, a default severity profile, a routing rule, and a resolution template. Categorization isn’t paperwork; it’s the basis for everything else the platform does with the exception.

PRODUCTION

QC failures, manufacturer delays, capacity shortfalls

When a lot fails QC, a factory misses its window, or capacity is overcommitted. Detected from production lot events and manufacturer status updates.

E-204 · Sapphire QC fail · 2 major defects vs. AQL 1.5 max 5 · 4h ago

SHIPMENT

Vessel delays, customs holds, transshipment misses

When a vessel runs late, a container clears customs late, or a transshipment is missed. Detected from carrier APIs, customs broker integrations, and route-deviation thresholds.

E-203 · M/V COSCO Pacific · +2 days · 6 orders affected · 12h ago

DOCUMENT

Missing, expired, or invalid documents

When a required document hasn’t been collected, a certificate is approaching expiry, or a generated document fails validation. Detected from the document chain on every order.

E-202 · Missing Certificate of Origin · RD-2026-077 · blocks customs · 8h ago

FINANCIAL

Cost overruns, FX shocks, margin compression

When a cost line accrues materially above estimate, an FX move impacts margin beyond threshold, or a payment goes past due. Detected from cost-line events and live margin computations.

E-198 · Athena quote margin · projected 19.8% vs. 22.5% target · 1d ago

COMPLIANCE

Sanctioned-party hits, restricted-substance flags, regulatory failures

When a counterparty matches a sanctions list, a product fails a destination-country compliance check, or a regulatory filing is rejected. Detected from real-time screening on every transaction.

E-201 · CE certificate expiring · Crescent · 47 days · 12 orders affected

COUNTERPARTY

Manufacturer non-response, broker errors, freight forwarder mistakes

When a counterparty fails to respond within a defined SLA, makes a recordable mistake, or stops engaging. Detected from communication-log gaps and SLA timers per relationship.

E-200 · Sapphire EUR.1 outstanding · 3 days · auto-reminder sent · 3d ago

DETECTION · THREE WAYS THE PLATFORM SEES THEM

Three signals. Cross-checked, deduplicated, real-time.

EDMA detects exceptions through three complementary signals — the platform’s own event stream (anything that happens in the system emits an event, and rules run against the stream), threshold rules configured per tenant (cost variance > X%, certificate expiry < Y days, response time > Z hours), and pattern detection that compares current behavior against historical baselines. A manufacturer averaging three days late shouldn’t trip a delay alert; one always on time and slipping two days should.

EVENT STREAM · live

Every system action emits an event

Order state changes, lot status updates, document arrivals, cost accruals, message receipts, payment events. The exception engine subscribes to all of them. Configured rules fire as events arrive — typically <2s latency from event to exception classification.

247 events/day across all sections · ~12 trigger exception rules

THRESHOLD RULES · per-tenant

Configurable boundaries on every metric

Cost line variance > 5% of estimate. Certificate expiry within 60 days. Manufacturer response time > 48 hours. Margin compression > 2pp from quote. Each tenant configures thresholds for their risk appetite and operating discipline.

26 default rules · ~40 typical per tenant after customization

PATTERN DETECTION · baseline-relative

Anomalies vs. counterparty’s own history

A manufacturer with 99% on-time delivery slipping by 3 days is more important than one with 85% slipping by 1. The platform learns each counterparty’s baseline (delivery time, defect rate, response time) and flags deviations relative to that baseline, not against a global standard.

Active baselines per manufacturer, per route, per product

Detected events are deduplicated across signals — a vessel delay flagged by both the carrier API and a route-deviation rule produces one exception, not two.

TRIAGE · COMPUTED SEVERITY, NOT GUT FEELING

A formula, not a feeling.

Severity isn’t a checkbox. EDMA computes a composite severity score for every exception from three inputs — value at risk (the dollars on the line), time-sensitivity (how quickly the exception compounds if not resolved), and downstream impact (how many other things are blocked by inaction). The score determines whose queue the exception lands in, what channels it triggers, and how prominently it ranks in the inbox.

Triage breakdown · E-204 · Sapphire QC fail

Severity HIGH · score 84/100

Value at risk$231,200Lot value at stake

Over $200K threshold for HIGH

Time sensitivity24hWindow before margin compression compounds

Under 48h threshold for HIGH

Downstream impact4Records blocked by inaction

RD-2026-077 (order)SHP-2026-088 (shipment)INV-2026-014 (invoice)EUR.1 (document)

Computed severity84HIGH

Routing: Liam Okafor (Operations owner).
Channels: in-app + push.
Inbox rank: 1 of 8.

Severity recomputes every time a downstream record changes. Rank updates live.

ROUTING · WHO, WHEN, VIA WHAT

The right person, the right channel, the right urgency.

Severity drives routing. HIGH-severity exceptions route immediately to the assigned owner via in-app notification, push, and (during quiet hours) email digest. MEDIUM exceptions route via in-app + email. LOW exceptions surface in the daily briefing only. Routing rules are configurable — escalation paths, channel preferences per role, quiet hours respected per user. The platform never spams; routing is a curated decision, not a broadcast.

Routing matrix · default rules · editable per tenant

Severity	In-app	Email digest	Push	SMS / phone
HIGH	instant	immediate alert	yes	on escalation
MEDIUM	instant	daily 7:30am	no	no
LOW	instant	weekly summary	no	no
Resolved	archive	summary only	no	no

Quiet hours respected

Vio Predoiu (CEO)push 08:00–22:00 local · SMS only for HIGH after 22:00

Sarah Chen (Manager)all channels 07:00–19:00 weekdays · in-app only weekends

Liam Okafor (Ops)all channels 24/7 (on-call rotation)

Routing rules are tenant-configurable. Defaults work out of the box for most trade operations; customization handles edge cases — multi-region operations, follow-the-sun support, on-call rotations.

RESOLUTION · TEMPLATES, NOT FREE-FORM

Common exceptions have known plays. Run them.

For each exception type, EDMA ships a default resolution template — the decision tree, the actions to take, the records to update, the notifications to send. Operators can customize templates per tenant. The result: when E-204 (Sapphire QC fail) arrives, the operator doesn’t think “what do I do?” — they pick from three pre-computed options, each showing its downstream impact, with the recommended one flagged based on pattern history.

Resolution template · QC inspection failure · default for medical-grade

Used in 47 resolutions over last 12 months · 91% within SLA

01 · DETECTQC fail eventInspection result FAIL · classify severity

→

02 · TRIAGECompute severityVAR × time × downstream

→

03 · DECIDEOperator decision3 options, impact preview

→

04 · EXECUTEApply decisionRework / reject / escalate

→

05 · UPDATECascade updatesOrder, shipment, mfr score

→

06 · AUDITWrite trailReason, decision, impact

What the template does automatically

Pulls the inspection report into the exception detail
Computes margin impact for each decision option (rework / reject / escalate)
Identifies the manufacturer’s recent pattern (3 of 4 similar resolved by rework)
Pre-fills the audit trail entry once decision is made

What requires the operator

Choosing among the three decision options
Confirming or overriding the recommended path
Adding context to the audit trail (optional, but enabled)

Templates ship with sensible defaults. Customizing them is a per-tenant Settings activity that takes about 20 minutes per template.

THE EXCEPTION LIFECYCLE · FROM DETECTION TO LEARNING

Six phases. Average 4 hours from detect to resolve. Audit trail end-to-end.

Every exception in EDMA passes through six phases — detected, triaged, routed, acknowledged, resolved, and integrated into pattern history. The platform tracks time spent in each phase, surfaces stuck exceptions, and reports MTTR (mean time to resolve) per category. The lifecycle is the unit of operational discipline; reducing it is the unit of operational improvement.

Lifecycle · E-204 Sapphire QC fail · in flight

MTTR THIS MONTH

4.2h

Across 47 resolutions. Down from 6.8 hours 12 months ago.

STUCK EXCEPTIONS

Currently exceeding SLA · oldest 18 hours past target. Auto-escalated to manager.

FULL TRAIL

100%

Every phase, every actor, every decision recorded. Read-only · queryable by category, owner, root cause.

PATTERN LEARNING · WHEN ONCE BECOMES THREE TIMES

Patterns are visible. The platform points at them.

A single Sapphire delay is an exception. Three Sapphire delays in four months is a pattern. The platform tracks every resolved exception against the counterparty, route, product, and root cause involved — and surfaces patterns when they cross thresholds. Patterns drive concrete actions: manufacturer score adjustments, dual-sourcing recommendations, route-pricing updates, and template tuning. Exceptions resolve. Patterns improve.

Detected patterns · last 90 days · ranked by impact

AllManufacturersRoutesProductsDocuments

Sapphire Industries · 3 delays in 4 months

Lots PL-2026-0098, PL-2026-0119, PL-2026-0142 · cumulative 11 days lost

RECOMMENDDual-source 30% of next quarter to Lotus Rubber Works

Impact~$380Kmargin protection

COSCO trans-Pacific lane · BAF surcharge frequency rising

4 surcharges in last 90 days vs. 1 in prior 90 days

RECOMMENDIncrease next-quote freight floor by $0.0002/unit

Impact~$48Knext-quarter margin

Certificate-of-origin generation · 6 stuck on EUR.1 from Crescent

Average wait 2.7 days vs. baseline 0.5 days · root: QA team capacity

RECOMMENDUpload checklist directly to supplier portal · auto-pull

Impact~2 daysper affected order

SafeHands UK · payment cycle extending

Last 4 invoices averaged 67 days vs. baseline 42 days

RECOMMENDDiscuss revised terms · flag in next review meeting

Impact~$210Kworking capital

Patterns close the loop between exceptions and operations. Resolved isn’t done; integrated into prevention is done.

VS. THE TOOLS YOU MIGHT BE USING

Built for trade exceptions, not IT incidents or support tickets.

Capability	Email + Excel	Zendesk	PagerDuty	TradeOS
Detection from system events (no manual reporting)	—	partial	✓	✓
Trade-specific exception taxonomy	—	—	—	✓
Computed severity (VAR × time × downstream)	—	—	partial	✓
Resolution templates with impact preview	—	partial	partial	✓
Cascade into orders, shipments, finance	—	—	—	✓
Pattern learning across counterparties	—	—	—	✓
Read-only audit trail across the lifecycle	—	partial	✓	✓

PagerDuty brings incident-management discipline to IT. Zendesk brings ticket-resolution workflow to customer support. Both are excellent at what they do. Neither knows what an Incoterm is, neither models a multi-manufacturer order, and neither closes the loop between a single exception and the next quarter’s pricing. EDMA does.

FREQUENTLY ASKED

Five questions about exception handling.

How is this different from the operational briefings cockpit?

The briefing is a curated daily summary — what changed, what needs you, what’s compressing margin. The exception inbox is the all-day workspace where exceptions actually get resolved. They share the same underlying event stream and severity scoring; the briefing surfaces the top items, the inbox holds the full queue with the resolution panel attached.

Can I configure my own exception types and detection rules?

What happens to an exception that no one resolves?

How are pattern thresholds set?

Does the platform make resolution decisions automatically?

See the exceptions you don’t know about, before they become problems.

Book a 30-min demo and we’ll set up your detection rules, your routing matrix, and walk through real exceptions on real workflow data — from detection to resolution to the pattern that drives the next quote.

Book a demo Or book a demo →