Source citations

Every value, with the page and line it came from.

InfoNotData never writes about a patient. It extracts specific values, like an ejection fraction or an ECG date, and links each one to the exact page and line in the original fax. Anyone can check it in one click.

Request a demo Extraction vs. generative AI

Extraction only
No generated text
Click through to the source

Illustration · synthetic record, not patient data

01 / Why citations

A fact you can’t check is a fact you can’t use.

Clinicians can’t act on a value they can’t trace. That’s why InfoNotData extracts and cites instead of summarizing.

Unverifiable

Summaries hide their sources

A paragraph about a patient doesn’t say where each claim came from. Checking it means reading the record anyway.

Invented

Generated text can hallucinate

A model that writes prose can produce a number that isn’t in the record at all.

Lost

Nuance gets flattened

A summary decides what matters. The specialist’s checklist should decide that instead.

02 / How it works

Extract the value. Keep the receipt.

Find
The checklist says what to look for: an ECG within 90 days, a troponin trend, an ejection fraction.
Only what the specialty needs.
Extract
The value is pulled out exactly as it appears. Nothing is paraphrased or generated.
The record’s words, not a model’s.
Cite
The value is linked to its page and line, like “Page 212, Line 14.”
The source travels with the fact.
Verify
One click opens the original page at that line, so anyone can check it.
Trust that can be audited.

03 / Extraction vs. summarization

Why InfoNotData doesn’t write summaries.

Summaries are useful for many things. Clinical intake isn’t one of them, because every value has to be checked.

Comparison of generative summaries and InfoNotData extraction
Aspect	Generative summary	InfoNotData extraction
Output	Paragraphs about the patient	Specific values from the record
Source for each claim	Usually not shown	Page and line for every value
Can it invent a value?	Yes, that’s the hallucination risk	No text is generated
Verifying it	Read the record to check the summary	One click to the cited line
When it isn’t sure	Often says something anyway	Routes the page to a person

Category comparison, not a comparison with any named product.

04 / When it isn’t sure

It says so.

What InfoNotData never does

Fill a missing value with a guess
Cite a page it didn’t read
Hide a low-confidence page inside an “auto-processed” result
Write text about a patient

Not found

Missing items are flagged

If the checklist asks for a stress test and it isn’t in the packet, the item is marked “not found,” so someone can request it.

Low confidence

Uncertain pages go to people

Handwriting or a poor-quality fax is routed as Hard and reviewed by a person on your team.

Questions about citations

What clinicians ask about citations.

What exactly is cited?

Every extracted value is linked to the page and line of the original document it came from, for example “Page 212, Line 14.”

Does InfoNotData ever generate text?

No. It is extraction-only. It never writes summaries or any other text about a patient.

What if the value appears in more than one place?

Each value carries the source it was extracted from, so a reviewer can always see exactly where it came from.

Can a reviewer see the original page?

Yes. One click opens the original page at the cited line.

06 / Next step

Check the citations yourself.

A 30-minute walkthrough with an engineer. We’ll extract values from a synthetic or de-identified packet and you can click through to every source.

What happens next

An engineer replies to set a time that works for your team.
A 30-minute walkthrough on a synthetic or de-identified packet.
If it fits, we scope a pilot on your own queue.

An engineer, not a sales script. We’ll show you what it doesn’t do, too.

Tampa, Florida

Where do your faxes land?

Organization type

Hospital or health system Community Care network Specialty group VA or federal health Post-acute, home health or lab Something else

Inbound fax volume Pages a month. A rough estimate is fine.

Who should we talk to?

Full name

Work email

Role (optional)

Organization

Anything we should prepare?

Notes (optional) For example: your specialty, your EHR, or your security review timeline.

Business contact details only. Please don’t include any patient information.