1. Why does moving to a new EMR take weeks?
Three jobs, all reading work, all done by hand. None of them is hard. Each is slow because every clinic’s files are a little different.
| The job | By hand | With AI in Aminova |
|---|---|---|
| Working out what each export file and column is | Someone opens every file and guesses what “pt_ref2” means | AI reads the file and column names and suggests a match, with its reason |
| Filing thousands of scanned documents | Everything lands in one folder, or someone sorts it page by page | AI reads page one and files labs, consents and insurance cards where they belong |
| Rebuilding intake questionnaires | Retyping every question into a form builder | AI turns each PDF into a draft online form |
| Redrawing consents for e-signature | Dragging a signature box onto every line of every PDF | One click places the signature, initials and date fields |
2. What are the steps of an AI-assisted import?
Six steps, and the clinic can run them itself. A signed-in owner or manager starts an import from Settings → Data & Privacy → Data imports. Progress is saved on the server, so you can close the window and pick it up tomorrow, on another computer.
| # | Step | What AI does | What you do |
|---|---|---|---|
| 1 | Add your files | Nothing yet | Upload the export: spreadsheets (CSV or Excel), ZIP files and the documents themselves |
| 2 | Reading your files | Works out what each file is and maps the columns it recognizes | Wait for it to finish |
| 3 | What we found | Shows suggested matches for the columns rules could not place, with a reason and a confidence | Approve, change or ignore each suggestion |
| 4 | Moving it into your clinic | Files documents into chart folders where it is confident | Nothing |
| 5 | Have a look around | Nothing | Open a few charts and check them against the old system |
| 6 | Import complete | Nothing | Confirm, then rebuild forms and consents |
If a file arrives in a format we cannot read yet, such as a raw database backup, the importer names it and tells you what to ask for instead, usually a CSV export. It does not quietly file it as a document.
3. How does AI read another EMR’s data export?
From the structure of the files, not the patients in them. To decide what a spreadsheet is — a patient list, a medication history, a document catalog — the AI is shown the file name and the column names. Not a row, not a sample, not a redacted example.
Our rules map the columns they recognise first. The AI is asked only about the ones left over, and each answer comes back as a suggestion with the AI’s own reason and a confidence of high, medium or low. A person approves it. Where the AI and the rules disagree about what a file is, the disagreement is shown to you rather than settled quietly.
Scanned documents are different: there is nothing in a file name that says whether a PDF is a lab, a consent or an insurance card. So the AI reads the first page only and files the document in the right chart folder, and it moves a document only when it is confident. Anything it is unsure of stays where you would expect to find it.
4. How does AI turn PDF intake forms into live forms?
Upload the PDF, and you get back a draft online form with the same questions in the same order. In the Intake Engine, Import from PDFs reads each PDF, or a whole folder of them, and produces one form per document, so a new-patient questionnaire, a hair-loss questionnaire and a thyroid history stay three forms rather than one long one.
| On the paper form | In the rebuilt form |
|---|---|
| Name, phone, email, date lines | Text, phone, email and date questions |
| A list of conditions to tick | A multiple-choice question with the same options |
| A single “I agree” box | A yes or no question, so it carries its own context |
| A signature or initials line | A signature question |
| A form written in Spanish | A form still in Spanish; it is not translated |
The rebuilt form is a draft. You edit wording, reorder sections or add the questions you always meant to ask, then publish it and send it before visits like any other form. Answers can then fill the matching chart fields.
5. How does AI place signature boxes on consent PDFs?
In the consent library, Read PDFs and place fields does it in one click. The AI finds every blank on the page and places the right field on it: signature, initials, date, text or checkbox. Each field is assigned to whoever signs it, the patient or a member of staff.
Two details matter more than they look. Date fields on a consent fill in with the signing date automatically, so a “Date of birth” line is always placed as text, never a date, or every patient’s birthday would be stamped with today. And every rebuilt consent is saved as an inactive draft: it is not sent to anyone until someone opens it in the PDF designer, checks the boxes and saves it.
Consents that came in with your import can be read the same way, straight from the import, so the agreements your patients already know move across with their signature lines in place.
6. What patient data does the AI see during an import?
As little as each job needs, and all of it under a signed BAA.
| Job | What the AI is shown | What it is never shown |
|---|---|---|
| Working out what a spreadsheet is | File name and column names | Any row or cell |
| Suggesting column matches | Column names left unmapped | Values in those columns |
| Filing a scanned document | The first page of that document | The rest of the chart |
| Rebuilding an intake form | The blank or sample form PDF you upload | Other patients’ answers |
| Placing consent fields | The consent PDF | Signed copies of other patients |
The AI runs on a HIPAA-covered model under a signed Business Associate Agreement, and under those terms what we send it is not used to train the model.
7. What should you get from your old EMR before you start?
A complete export in standard formats, and the original PDFs of every form and consent you use. The AI is fast with whatever you bring; it cannot recover what was never exported.
| Bring | Why |
|---|---|
| Patient demographics, as CSV or Excel | The roster everything else attaches to |
| Clinical notes, medications, allergies and labs | The chart history your providers rely on |
| Documents and scans, as a ZIP | Labs, consents and insurance cards the AI will file |
| Each intake questionnaire as a PDF | To rebuild as live forms |
| Each consent as a PDF | To place signature fields |
Our switching guides cover what specific systems export and how: Cerbo, OptiMantra, Healthie, Vagaro and Aesthetic Record. For what the law says about getting your data out, see what happens to your data when you switch.
8. Frequently asked questions
How long does it take to switch to an AI EMR?
Our own onboarding went from about three weeks to about three days once AI took over the first pass of mapping, form rebuilding and consent fields. Your timeline also depends on how quickly your old system produces its export.
Can AI migrate data from my old EMR on its own?
It does the reading, and a person approves the result. In Aminova the AI suggests what each file and column is; nothing it suggests is applied until someone accepts it.
Does the AI see my patients’ data during the import?
To understand a spreadsheet it sees only the file and column names. To file a scanned document it reads that document’s first page. It runs under a signed BAA, and the data is not used to train the model.
Can AI convert my PDF intake forms into online forms?
Yes. Upload the PDF and Aminova returns a draft online form with the same questions in the same order, one form per PDF, in the form’s own language. You edit it before sending it.
How do I add signature boxes to a consent PDF?
In Aminova’s consent library, Read PDFs and place fields places signature, initials, date, text and checkbox fields on the lines in one click. You check them in the PDF designer before the consent goes live.