AI Document Intelligence
The Problem
A fax arrives as pages with nothing on them to say what they hold, and nothing moves until a person opens it and reads it. Referrals are a small share of what arrives, so the workload is set by page count rather than by the number of patients waiting to be seen.
Referrals sit behind the rest.
Nothing moves until someone opens the fax and recognizes what’s inside.
01
Documents arrive unlabeled.
Coordinators split and file each one by hand before the chart is readable.
02
A filed document tells no one.
It can reach the chart and still never reach the person who needs to act.
03
Hatch reads every page as it arrives, finds where one document ends and the next begins, and identifies each one by type. A single fax can hold several documents and more than one patient.
Patient, insurance, referring organization, and reason are extracted into fields, and the patient match is proposed for a person to approve. A reviewer confirms and splits each document out with one action.
Every document type gets a destination your clinical teams already use, whether that’s a team inbox, a shared drive, or a task in the EHR. Hatch notifies the person who needs to act.
The Benefits
Full visibility
You can see who works on what, and how far behind you are. A new hire does not have to learn which documents matter.
Physician readiness
You find the MRI without scrolling through a 40-page file. Hatch notifies you when a new document lands in the chart.
Human-verified accuracy
Your team looks at every document at first. Once it trusts the results, it looks only at the exceptions. Every time your team approves or corrects a document, Hatch learns your rules.
What types of documents can Hatch process?
Referral orders, imaging reports, lab results, medical records, insurance and authorization documents, and demographic forms. The document types are defined with the group during implementation and match the labels the organization already uses. If it arrives by fax, email, portal, or direct upload, Hatch reads it.
What if our EHR isn’t integrated with Hatch?
Most of the work still runs. Hatch reads every page, separates the referrals, splits and labels the documents, extracts the data, and routes each document to the destination the team uses. What the integration adds is chart creation and filing to the chart under the correct category, plus patient matching against the EHR record. Without it, a person uploads the document to the chart, with the patient details shown alongside so nothing has to be looked up.
How accurate is the extraction?
Every document is confirmed by a person before it is filed, and the patient match is proposed rather than applied. Teams look at everything at first and shift to reviewing only the exceptions once they trust the results. Corrections train the system on the group’s own rules, so accuracy improves against the documents that organization actually receives.
Can Hatch handle a fax with mixed content or more than one patient?
Yes, and both are common. A single fax can carry a referral order, progress notes, labs, and an insurance card, and it can carry records for several patients or for none of the patients it appears to be about. Hatch reads the full document, groups the pages into separate documents, and lets the team split each one out and assign it to the correct chart.
Where does each document go once it’s identified?
To one destination per document type, set by the group during implementation. That can be a team inbox, a shared drive, an email address, or a task in the EHR where the EHR supports it. Hatch notifies the person responsible, which is the step that filing alone does not accomplish.
Our fax numbers are already routed into our EHR. How does Hatch receive them?
Three paths, chosen during implementation. The referral fax lines can be ported to Hatch, an existing inbox can auto-forward to Hatch, or documents can be pulled through the EHR API. Referring organizations are not asked to change anything.
Where do our documents actually live?
In the EHR. Hatch mirrors and pushes data into it and does not pull records out to store elsewhere, so documents and clinical data stay where compliance and continuity require them to be. The EHR remains the system of record.