AI SmartDoc — What It Does and How It Works | AI SmartDoc

AI SmartDoc is Remindax's automated document intake. You upload a document (PDF, DOCX, JPG, PNG), SmartDoc extracts the key fields — title, expiry date, document type, plus relevant custom fields — and pre-fills a new reminder for you. Instead of typing every field for every renewal, SmartDoc gets the data off the page. Documents are processed by OpenAI as part of the extraction.

What SmartDoc is good for

SmartDoc shines on documents that arrive routinely with predictable fields:

Insurance certificates — expiry date, policy number, insurer name
Licenses and permits — issue date, expiry, issuing authority
Vehicle registrations — plate, VIN, expiry
Trade licenses — license number, expiry, holder name
Contracts — start date, end date, renewal terms

If you process 5+ documents in a session, SmartDoc removes meaningful manual data entry. For one-off documents, manual reminder creation may be faster.

What SmartDoc extracts

When you upload, SmartDoc tries to find:

Title — a meaningful name
Expiration / due date — auto-populated into the reminder
Document type — matched against your workspace's existing types (or proposed as a new type to create)
Notes — relevant context from the document
Custom field values — where SmartDoc recognises fields you've defined
Tags — where relevant

After SmartDoc finishes, you arrive at the normal new reminder form with these fields pre-filled and an extraction summary above. Review, adjust, Apply, then Save.

What SmartDoc doesn't do

It doesn't read everything perfectly. Handwriting, low-resolution scans, unusual layouts, and multi-language documents reduce accuracy. Always verify before saving.
It doesn't notify recipients. SmartDoc creates the reminder; notifications follow your normal sequence rules.
It doesn't replace your judgment on document type. SmartDoc proposes; you confirm.
It doesn't analyse legal contracts. SmartDoc extracts identifying fields, not legal terms or obligations.
It doesn't attach to existing reminders. SmartDoc always creates a new reminder — there's no "scan a renewal and attach it to the existing item" workflow.
It doesn't split bundled PDFs. A PDF containing multiple unrelated documents is treated as one. Split the file first.

How it works (the technical picture)

You upload a document via AI SmartDoc → Scan with AI.
PDFs and images are sent to OpenAI for field extraction.
DOCX files are read locally first to extract text, which is then sent to OpenAI.
The file itself is stored in Drive in your workspace's AI SmartDoc folder, on AWS storage.
Extracted fields populate the new reminder form, with a summary above.
You review, Apply, and save — Remindax creates the reminder with the data and the file attached.

Third-party processing disclosure

The current SmartDoc flow sends document content to OpenAI for extraction. Older OCR code in Remindax also touches AWS Textract in some paths, but the SmartDoc UI you use today routes through OpenAI.

Per OpenAI's stated API policy, API inputs and outputs are not used to train OpenAI's models by default unless the organization opts in. See OpenAI's data usage policy and OpenAI's business data privacy for OpenAI's own terms.

If your data classification policy prohibits sending document content to third-party AI services, don't use SmartDoc for that content. Upload files directly through Drive's standard upload, which stores the file in AWS but doesn't invoke SmartDoc's extraction. See Where your data is stored and who processes it.

Usage limits

SmartDoc enforces a per-user, per-month token usage cap based on your plan:

Plan ID 1 (entry plan): 20,000 tokens per user per month.
Other plans: 300,000 tokens per user per month.

When you hit the limit, the SmartDoc page shows a warning and further scans are blocked until the cap resets at the start of the next month.

Token usage is the input + output your document generates with OpenAI — a short, clean document uses fewer tokens than a long, dense one. Most teams don't hit the cap; if you do, it's a signal to either upgrade or to split SmartDoc usage across more accounts.

Supported file types and sizes

File types: PDF, DOCX, JPG, PNG
Max file size: 25 MB per document
One document per scan

For larger or different file types, upload directly to Drive (no SmartDoc), or extract the relevant pages into a smaller PDF first.

What happens next

After Scan with AI finishes, you land on the new reminder form with fields filled.
Review every field — especially the expiry date.
Click Apply to confirm the SmartDoc result, then Save to create the reminder.
The file is saved to your AI SmartDoc folder in Drive, linked to the new reminder.

Edge cases & gotchas

No side-by-side preview. Review uses the normal reminder form layout — the extracted fields appear in the form, with the document accessible for reference.
Scan quality matters. A clean 200-DPI scan extracts far better than a glare-affected phone photo.
Multiple expiry dates on one document. SmartDoc usually picks the most prominent date. If that's wrong, adjust before applying.
Documents with no expiry date. SmartDoc may guess based on layout heuristics. Always confirm.
Foreign-language documents. Accuracy varies; review extra carefully.
Bundled PDFs are treated as one document — split before uploading.
Token limits reset monthly per user. Hit the cap and you're blocked until next month or until you upgrade.
Don't use SmartDoc for highly sensitive content if your data policy forbids third-party AI processing.

AI SmartDoc — What It Does and How It Works