The Answer Was at a Logistics Show.

We wanted to return a task from two people to one without pushing safety back onto attention alone.

01 / THREE PRIORITIES

Accuracy first, speed second, low cost third.

Two-person verification improved accuracy but consumed more labor. Returning to one person for speed or cost could bring the same errors back. None of the three priorities could simply be discarded.

PRIORITY 01AccuratePrevent mix-ups and missing specimens.

PRIORITY 02FastDo not delay the start of testing.

PRIORITY 03SustainableKeep the operation affordable.

02 / ENCOUNTER

The answer was in the exhibition next door.

While visiting an exhibition at Tokyo Big Sight, I noticed a logistics show taking place alongside it. A warehouse case study used voice picking to reduce a two-person task to one person.

Specimen handling is also a kind of picking: select the correct physical item for an order and pass it safely to the next process. I thought, “If we bring this idea into specimen verification, perhaps one person can do it here too.”

03 / TWO-WAY VOICE

The computer does more than speak.

The warehouse system did not merely read instructions aloud. Speech synthesis delivered the instruction; when the worker replied, speech recognition received the answer.

In our laboratory, the system read the patient name and required tube. The worker selected the tube and answered “Yes” or “Next.” The application recognized the response and advanced to the next check.

SYNTHESIZEThe system speaksRead the patient and required tube aloud.

ACTThe person selectsListen and remove the correct physical tube.

RECOGNIZEThe system listensRecognize “Yes” or “Next” and continue.

We did not replace only the “reader.” We designed a partnership in which the application communicated and received the worker’s intent.

04 / THE QR ALREADY THERE

Separate from voice, every tube already had a QR code.

A barcode reader was already in use. Each tube had a unique QR label, and the scanned identity was matched with the order shown on screen. The QR data also carried the tube type.

VOICEDialogue with the application

The system speaks and recognizes the person’s response.

SPECIMEN QRIdentify the physical item

Confirm exactly which tube has just been scanned.

At this point, QR codes only identified specimens. Then, while watching people’s hands, I imagined a second job for the same reader.

05 / RESULT

One person—and fewer mistakes.

A computer says the same information consistently. It does not misspeak or get tired, and it receives the worker’s answer directly. The person can concentrate on selecting the correct tube.

People assigned to verification

2 → 1Roles redesigned without returning accuracy to attention alone

I did not search for a technology.
I searched for similar work.
A logistics idea that could speak and listen became a safer clinical-laboratory process.

CHAPTER 02 · EPISODE 02 / 09

Chapter 2 index: 9 stories + extra