Tell Davner what the chart doesn’t know
Staff can now add notes before Davner reads a case, giving our extraction process the context needed to get more requests right the first time.


Davner handles prior authorizations from beginning to end. To get started, practices give us the patient's clinical records through an EHR integration, an electronic health information (EHI) export, or a direct upload of PDFs and images. We pull out the patient, insurance, procedure, provider, and clinical information required for the authorization and present it for a quick review.
The point is to avoid making staff fill out another long form before we go fill out the payer's long form. That would only be marginally better than doing the authorization themselves.
But the patient's chart doesn't always contain everything the practice knows.
Some information lives in an email, on a piece of paper, or in a conversation with the patient. The surgery date may have changed. A patient may want their MRI performed at a facility closer to home instead of the facility the practice normally uses. The doctor may operate at one location every Tuesday and Wednesday except this week, when the schedule is different.
Those exceptions can't always be turned into a permanent rule, and requiring staff to update the chart and upload it again creates more work for a basic fact.
Staff can now include notes or special instructions when they upload a case to Davner. They can type a sentence, paste an email, dictate the information, or use the field however they see fit. Davner reads the note together with the clinical records before extracting the information for the authorization.
Get the exception right the first time
Previously, staff could submit the records and correct the extracted information during the review step. They could also add a comment afterward if the information didn't fit adequately into the form.
That worked, but it required them to remember the exception after the case had already been processed. The extra step becomes even more annoying when several cases are being extracted at the same time.
Moving the note before extraction gives the AI the missing context upfront. If a patient changed imaging facilities after the encounter, for example, staff can say so while uploading the records. Davner can use the new facility, look it up in the NPI registry, and fill in the corresponding information during the initial extraction instead of asking the staff member to replace it later.
The same applies to a new date of service, corrected member ID, procedure code, payer detail, or almost anything else the staff knows that the documents don't. If a provider ordered several procedures but only a subset should be authorized right now, staff can explain that too instead of removing the other orders from the record or correcting the request later.
Flexible doesn't mean unchecked
When a staff note fills a gap or disagrees with the uploaded documents, Davner prefers the staff-provided information. That's important because the note often exists specifically to explain why the record is no longer current.
But typos are a fact of life. Structured information such as clinical codes, provider names, and NPIs can be checked against the appropriate databases. If a staff member enters a nearly correct diagnosis code or provider detail, Davner can identify the mismatch instead of silently building it into the request.
The extracted case still goes through review, and the submitter has the final say. We also preserve the original note in the authorization history and maintain a human review process on our side. That creates a deliberate check against AI mistakes and against automation fatigue, where someone gets used to clicking approve without looking closely.
An escape hatch for the real world
We could have responded to every exception by adding another field to the intake form. That would make the form longer while still failing to cover whatever unusual situation comes next.
The notes field is an escape hatch. It makes the extraction process more flexible without requiring the practice to rebuild its workflow or translate every piece of knowledge into a structured form.
Davner was co-founded by a practicing physician. We live in the real world, where information isn't always complete, schedules change, and the record isn't always perfect on the first try. This gives staff one more easy way to tell us what we should know so we can get the case right on the first pass, with fewer corrections and fewer opportunities for mistakes later.
Notes and special instructions are live now when uploading prior authorization requests.
