SprigganSpriggan

Product · case study

What happens to a fax at 9:41 AM

Aug 25, 2026 · 7 min read · Spriggan team

All posts

01The problem

The fax machine is still how medicine talks to itself. Imaging centers, labs, specialists and payers all send paper, and in a busy primary care clinic that paper arrives faster than anyone can read it. A medical assistant we worked with kept a physical tray on her desk and a second one under it for the days she fell behind.

The cost is not the filing. It is what waits in the tray: an abnormal potassium, a denied authorization, a records request with a deadline. Every one of those is invisible until a person opens the page, decides what it is, finds the patient, and tells the right physician.

02What we saw

We timed it. From arrival to being filed in the correct chart, a typical fax took between forty minutes and the next morning, depending on the day. Roughly one in ten was filed to the wrong patient or the wrong section at first pass, usually because two patients shared a surname or the cover sheet was missing.

Nobody was careless. The work was simply unreadable by the software they had, so it was done by the people who were also answering the phone.

Fax · 3 pages · Bayview ImagingreadMarcus BellChart 5821 · DOB 03/12/71Matched · name, DOB, payer IDFiled under ImagingSummary added to the visit note · 4 s after receipt
A fax being read: three pages scanned, the patient matched on name, date of birth and payer ID, the document typed as an imaging report.

03How Spriggan handles it

At 9:41 a fax arrives from an imaging center. Before it finishes printing, the agent has read all three pages, found the patient from the name, date of birth and payer ID, matched them to the chart, and classified the document. An imaging report, not a records request or an authorization decision. By 9:41 and four seconds it is filed under Imaging in the right chart, with a one-line summary on the visit.

The same reading sorts the whole inbox. Lab reports go to results, authorization decisions to the prior-auth agent, records requests to staff, referrals to the referral queue. Each clinic sets its own rules; the agent applies them to every page, every time.

Fax inboxLab report · QuestResultsAuth decision · AetnaPrior Auth agentRecords requestStaff inboxReferral · Dr. OkaforReferrals
The inbox sorting itself: each document gets a destination and leaves the pile.

04What AI made possible

Then it reads the report the way a clinician would. Two values are out of range. Not critical, but the physician should know today, not on Friday. A text goes out with the two numbers and a link to the full report. Nine minutes later the reply comes back: recheck in three months, add it to the plan. That order is placed from the reply.

This is the part that was not possible before. Rule-based systems could route a fax by sender. They could not tell that page two of an unstructured report contains a value the physician cares about, or that the physician’s three-word reply is an order. Reading, in both directions, is what turns a filing system into a colleague.

Lab report · QuestHbA1c7.9%HLDL128 mg/dLHCreatinine0.9 mg/dLAbnormal · not critical · notifyMarcus Bell: HbA1c 7.9% (H), LDL 128 (H).Full report is in the chart.Seen. Recheck A1c in 3 months,add to plan.Order added to the chart · 9:44 AM
The abnormal values, the text to the physician, and the reply that becomes an order in the chart.

05What changed

The tray is gone. Faxes are filed in seconds instead of hours, to the right chart, with the original pages attached and searchable. Abnormal results reach the physician the same morning, on the phone, with the numbers in front of her. Orders and referrals go back out by fax straight from the chart, with the confirmation filed where the visit lives.

Nobody touched the fax. Nobody re-typed a value. One person made one decision. That is the whole product, in one document.

The fax was in the chart before I knew we had received it. The first I saw of it was the two numbers on my phone.

Share this case study