Automated job tracking: know where every order is

Why "where is my order?" is so expensive

In most mid-size shops, the status of a job lives in people's heads, on a whiteboard, or in a spreadsheet somebody updates when they remember. When a customer asks where their order is, someone walks the floor, calls the shop lead, or logs into a carrier website. The answer is right, but it costs a person's time every time, and the customer waits.

The fix isn't asking people to type more status updates. It's using the signals the business already creates.

Many signals, one record per job

Nobody types a status. Each signal the business already produces moves the job forward.

Peekgentic diagram.

The four signals you already have

What each signal can tell you

	Order email	Barcode or QR scan	Photo	Carrier tracking
Tells you	A job exists: customer, items, dates	Exactly where a part or box is	A step happened: inspected, packed, loaded	Where a shipment is, and when it should arrive
Needs	Software that reads the order	Labels and a scanner at each step	The photos your team already takes, and a vision model to read them	The tracking number tied to the job
Effort to add	Low	Medium: labels, scanners, habits	Low: no new hardware	Low

Peekgentic summary.

1. Order emails

The job starts when the order arrives. If software reads each order from email and PDF and creates the job record, every later signal has something to attach to. That's the foundation. Our guide on automating order entry covers this step.

2. Barcode and QR scans

Barcodes are the most precise signal: a scan says exactly which item is where, and when. According to GS1 US, the standards body for barcodes, they work by encoding numbers into a pattern of bars, dots or boxes, and 2D codes like GS1 DataMatrix can hold more information than the classic lines. The cost is in the habit: someone has to scan at each step.

3. Photos

Many shops already photograph travelers, finished units and packed pallets. A vision model can read the job number, label or serial plate in each photo and mark the step done, with no new hardware and no extra clicks. It isn't perfect: blurry, rotated or tiny text causes mistakes, so anything unclear goes to a person. Our guide to vision AI on the factory floor covers what these models can and can't read.

4. Carrier tracking

Once a job ships, the carrier already tracks it. Carriers offer this to businesses directly. FedEx, for example, offers a tracking API that returns a shipment's status, scan events and estimated delivery window. Tie the tracking number to the job, and the record updates itself until delivery.

What good tracking looks like

• One record per job, not one per system. The order, the floor, the photos and the shipment all point to the same job.
• Status nobody types. Each step is marked by a signal, with the evidence attached: the scan, the photo, the carrier event.
• Exceptions, not dashboards. The useful alert is "this job hasn't moved in three days," not a chart nobody opens.
• Anyone can answer the customer. Sales, the front desk and the owner see the same status.
• A person handles the unclear ones. A photo the model can't read, or a scan that doesn't match an order, goes to someone, not into the record as a guess.

How to start

1. Start where the job starts: make sure every order creates a job record automatically.
2. Add the cheapest signal next: usually carrier tracking, then the photos you already take.
3. Add scans only where precision pays, such as parts that get lost or mixed up.
4. Measure one thing: how long it takes to answer "where is my order?" before and after.

This is the kind of system we build around the tools a company already uses. See order processing and photos and inspections.