The Work Order Every Tech Ignores (and Why It Is Not Their Fault)

Pick up a work order from any leading lab management system and look at what the technician actually sees. Product codes, alloy charges, shipping line items, barcodes, customer preferences formatted in dense bracket notation, all on one page with no visual hierarchy. The document reads like a system report, not an instruction sheet.

A technician at the bench needs five things: tooth number, shade, material, restoration type, and any special instructions. Everything else is operational data that belongs in the back office, not on the production floor. When it's all presented together with no distinction, the critical details disappear into the noise.

The result is consistent across labs. Techs glance at the shade, confirm the tooth number, and work from memory. The work order gets clipped to the case pan and functionally ignored. Not because the tech is careless, but because parsing the document takes longer than remembering what needs to be done. When 90% of cases are standard, the eyes learn to scan for the standard information and skip the rest.

That's how special instructions get missed. A note in the margin ("raise occlusion 1mm," "tight mesial contact") doesn't register when the document around it trained the reader to stop reading.

---

The problem starts before the work order exists

The work order is the output. The input is whatever the dentist sent, and that changes every time.

Prescriptions arrive as digital scans, printed forms, handwritten notes, emails, and phone calls. The person at intake is usually office staff, not a trained technician. They're recording what arrived and creating the case in the system. Across four or five different prescription formats per day, each one puts critical information in a different spot, in a different format, with different terminology.

This is where errors start. Not because the staff is careless, but because the process asks them to interpret clinical documents they weren't trained to read. A handwritten prescription with ambiguous shorthand, a checkbox form where the dentist marked two conflicting options, an email that says "same as last time" with no reference number. Intake staff do their best to translate what arrived into the fields the system requires. Sometimes they get it right. Sometimes "make it longer" becomes "make it shorter," and nobody catches it until the case ships.

The work order inherits whatever intake captured. If the intake was clean, the work order is clean. If the intake involved guesswork, the work order carries that guesswork forward: now printed, official-looking, and clipped to a case pan where a technician will glance at the shade and start working.

So you have two problems compounding each other. The information going into the work order is inconsistent, and the information coming out is unreadable.

---

How we solved it

We addressed both sides: what goes in and what comes out.

On the input side, we removed the interpretation step. Instead of asking intake staff to read a handwritten prescription and manually key each field into the system, the system reads the prescription using Anthropic's Claude. A photo of a physical prescription gets scanned and parsed: tooth numbers, shade, material, restoration type, special instructions. Claude extracts each field and scores it by confidence. High-confidence fields are pre-filled. Low-confidence fields get flagged for review, with the original image displayed alongside the extracted data so the reviewer can confirm or correct in one pass.

The same logic applies to every format. An email from a dentist gets uploaded and parsed the same way. A phone call with additional instructions gets captured through Claude-powered speech-to-text and attached directly to the case, structured, visible to every tech who touches it, not living on a sticky note that may or may not make it to the bench.

The person at intake goes from interpreting prescriptions to confirming extractions. That's a fundamentally different task, and one that doesn't require clinical training to do well.

On the output side, we stripped the work order down to what the technician actually needs. Five fields by default: tooth number, shade, material, restoration type, special instructions, and due date. That's the view at the bench. No product codes, no billing data, no preference records going back three years. Every field that made the cut earned its place. Everything else was removed, not because it doesn't exist in the system, but because it doesn't belong on the document a tech reads at 7am.

Special instructions are surfaced as visual flags, not buried in a notes field. "Call doctor before starting." "Photos on file. Review first." "Rush case." These aren't hidden in free text. They're the first thing the tech sees after the case header. If there are no special instructions, the tech knows it's standard and moves on. If there are, they're impossible to miss.

The result is the same clean format every time, regardless of whether the prescription arrived as a photo, an email, a phone call, or a digital scan. One consistent work order. One reading experience for the tech.

---

How you can do this right now

You don't need software to fix your work orders. You need to separate what the technician reads from what the office tracks, and you need to put special instructions where they can't be skipped.

Most work order templates fail because they're designed for one audience. Either they're built for intake staff (structured fields, lots of tracking data) and the tech can't find what matters, or they're simplified for the tech and intake has nowhere to put operational details. The fix is two zones on the same document.

Zone 1 is the tech view. It goes on top.

This is the only part a technician needs to read. It answers five questions: what am I making, what shade, what material, any special instructions, and when is it due.

Special instructions sit directly below the case header, above restoration details. This is deliberate. If instructions are below everything else, the tech reads the shade, confirms the tooth number, and starts working. They never scroll down. Put the flag second, right after the header. The tech initials line is the gate: you read it, you initial it, then you work.

Zone 2 is the intake and tracking view. It goes below.

The tech doesn't need to read this section. It's the operational layer: what was received, what's missing, and who signed off at each production stage.

Set this up as a Google Sheet. One sheet per case. Duplicate a template tab for each new case. Use dropdown menus for restoration type, material, and shade source. Intake staff shouldn't be typing "zirconia" freehand when a dropdown prevents the typo. Lock the label cells so nobody accidentally overwrites the structure.

Give the special instructions section a colored background, yellow or orange. When printed or viewed on screen, it stands out even from across the room. The visual weight matches the operational weight.

---

The person filling out the work order and the person reading it have completely different needs. Most systems pretend they're the same audience. They're not. The intake person needs structured fields to capture everything that arrived. The technician needs to glance at one page and know exactly what they're building, what shade, and whether there's anything unusual about this case. When those two needs live on the same document without separation, the tech stops reading, and the work order becomes decoration.

Separate the two views, surface special instructions where they can't be missed, and keep the tech-facing section down to the fields that actually matter at the bench. That's the entire fix.