Shipment Exception Report with Claude Cowork

What this workflow does

Use Cowork to group duplicate records, identify late or missing milestones, rank issues using documented rules, and draft owners and next checkpoints. It should not update dispatch, infer an ETA, or contact customers. This workflow helps operations teams transform raw TMS data into a prioritized exception list that humans can act on quickly.

The workflow is most useful when your TMS export includes hundreds of shipments and manual exception review takes hours. Cowork reads the data against your defined rules, flags exceptions, and ranks them by severity so operations can focus on the highest-impact issues first.

When to use this workflow

Use this workflow daily or weekly to stay on top of exception volume. Define your rules once (e.g., "late = more than 2 hours past promised delivery time"; "at-risk = dwell time exceeds 4 hours without pickup confirmation") and reuse them. Ideal for companies processing 100+ shipments daily.

Runbook

1. Freeze a time-stamped TMS or spreadsheet export.
2. Define late, at-risk, dwell, and data-quality rules.
3. Create one traceable row per exception with source row.
4. Operations verifies high-severity items in the system of record.

Real-world example

A parcel carrier pulls a TMS export of 500 active shipments at 8 AM. They define rules: "late = missed delivery window by 2+ hours"; "at-risk = no pickup scan in 6 hours"; "data quality issue = missing weight or signature". Cowork processes the export, identifies 47 exceptions (12 late, 18 at-risk, 17 data issues), and ranks them by severity. Operations reviews the high-severity list, contacts 3 customers proactively about late shipments, and assigns 2 data quality issues to dispatch for resolution. The process takes 15 minutes instead of 2 hours.

Copy-and-run prompt

Defining and refining exception rules

The core of this workflow is accuracy in exception detection. "Late" means something different at 8 AM on a Monday than it does at 9 PM on Friday. Define your exception rules carefully and update them as your business changes. For example: a parcel carrier might define "late" as "missed the promised delivery window by 2+ hours" on standard routes, but "missed by 4+ hours" for remote areas with inherent variability. Cowork will apply the rules you give it consistently; your job is to ensure the rules match your actual service levels and customer commitments.

When you first implement this workflow, plan to run it daily or weekly for a few weeks. This validation period helps you catch rule issues early. Common rule problems: time zone mismatches (your TMS is in UTC but delivery windows are in local time), missing milestone fields (you forgot to include a scan timestamp in the export), or rules that are too strict (flagging shipments as late when they're actually on track because your TMS hasn't updated yet). Once you've resolved these issues, the workflow becomes more reliable and you can trust the output more quickly.

Using exception reports for continuous improvement

Exception reports are not just for daily operations; they're also data for continuous improvement. Track exception trends over time: Are the same routes consistently late? Are certain carriers missing pickup confirmations? Are data quality issues concentrated in specific fields? Over a few weeks of reports, patterns emerge that tell you where to invest: better route planning, carrier performance reviews, or TMS data entry improvements.

Share exception reports with your performance management team, not just operations. If you're seeing high at-risk rates, that's a signal to hire another driver or negotiate better pickup windows with key customers. If you're consistently flagging data quality issues in a specific field, talk to your TMS vendor about adding data validation. This workflow generates the raw material for continuous improvement; your leadership team decides what to do with the insights.

Common pitfalls

Don't: Ask Cowork to update dispatch, modify routing, or change ETAs; keep those manual. Don't: Invent rules that require predictive logic (Cowork can't forecast delays). Don't: Rely solely on this workflow for critical issues; operations should verify high-severity items in the system of record. Do: Update your rules as business logic changes. Do: Run the workflow on the same schedule (e.g., every morning at 8 AM) so operations can rely on it. Do: Archive 90 days of exception reports so you can look for trends and prove to customers you're managing quality.

Human review checklist

• Verify every material fact against its source.
• Keep consequential actions manual.
• Flag missing, stale, or conflicting data.
• Use minimum permissions and approved files only.
• Check high-severity exceptions in the TMS before escalating to customers or drivers.
• Review rules quarterly; update thresholds as network performance changes.