Solutions
Turn faxes into structured Rxs
Built for the reality of smudged scans, handwritten corrections, multi-drug faxes, and peak-hour volume.
- Live in 2 days
- PHIPA compliant
- Canadian data residency
How fax intake works
Each document is assigned a unique intake ID on arrival. Multi-page faxes are split into individual prescription units where possible. Duplicate detection prevents the same fax from being processed twice, even if your fax provider retries delivery.
For the end-to-end fax-to-Kroll workflow and a pharmacist's view of the rollout, see the Kroll fax workflow guide and the pharmacist's guide to fax automation.
What the AI reads
The parsing pipeline runs in layers, not a single OCR pass.
Image normalization
De-skew, contrast adjustment, and noise reduction before any text extraction.
Field extraction
Patient name, DOB, provincial health number, prescriber, drug, strength, form, sig, quantity, and refills, identified by position and context, not pattern matching alone.
Sig interpretation
Shorthand codes, Latin abbreviations (QID, PRN, HS), and freehand variations are resolved to plain-language directions before Kroll mapping.
Confidence scoring
Each field gets a confidence level. Fields below threshold route to review rather than proceeding to write.
What AutoRx handles that OCR alone cannot
Handwritten Rxs
Full or partial handwritten prescriptions are parsed by the vision agent, not brittle text extraction.
Multi-drug faxes
Multiple prescriptions on a single page are split into individual processing units automatically.
French-language Rxs
French prescription formats and terminology are handled natively.
Non-standard forms
Pre-printed forms with variable completion quality are handled regardless of how they were filled.
Handwritten corrections
Corrections written over typed text are identified and the corrected value is used.
Prescriber shorthand
Non-standard sig notation specific to individual prescribers is resolved per source.
Volume and reliability
Fax intake is queue-based, AutoRx processes incoming documents in parallel, not sequentially. A Monday-morning surge of 200 faxes processes at the same throughput as a slow Tuesday afternoon. Kroll write failures retry automatically with exponential backoff. Permanent failures surface immediately in the dashboard with the original document attached and a clear failure reason. Your team resolves most exceptions in under 60 seconds.
Frequently asked questions
- What fax formats can AutoRx read?
- AutoRx accepts raw TIFF and PDF faxes, multi-page documents, and batch transmissions from your existing fax-to-cloud provider. It handles handwritten prescriptions, multi-drug faxes, French-language forms, and smudged or non-standard scans, because parsing runs as a multi-agent vision pipeline rather than a single OCR pass.
- Do I have to change my pharmacy's fax number?
- No. Your existing fax number stays the same. You point delivery to AutoRx's intake endpoint through your fax-to-cloud provider, and every inbound fax is forwarded into the AutoRx queue automatically. If you do not have a fax-to-cloud provider yet, AutoRx recommends one during onboarding.
- What happens when AutoRx is not confident about a field?
- Each extracted field gets a confidence score. Anything below threshold routes to your exception queue with the original fax attached and the uncertain field flagged, rather than being written to Kroll. A pharmacist or technician confirms or corrects it, and most exceptions resolve in under 60 seconds.
- Can AutoRx handle a Monday-morning surge of faxes?
- Yes. Fax intake is queue-based and processed in parallel, so a surge of 200 faxes processes at the same throughput as a slow afternoon. Kroll write failures retry automatically with exponential backoff, and permanent failures surface in the dashboard with the original document and a clear reason.
Get started with AutoRx
Ready to automate?
Most pharmacies go live within 2 days. No waitlist, and your existing Kroll setup stays exactly as-is.
Usually responds same business day