Outsource
Document processing outsourcing for scanned and digital files
Outsource document processing: sort, split, index and key documents, and check OCR and IDP output in your tools. $5 per person-hour, QA sampling included.
$5 per person-hour for standard tasks.
Document processing outsourcing hands the steps between "a file arrived" and "the data is in our system" to a trained team: sorting, splitting, indexing, extraction and exception handling. Deepen AI does this work in your document tools, alongside whatever OCR or intelligent document processing (IDP) software you already run.
What the work is, and who buys it
Buyers are finance, operations, insurance and logistics teams with a steady inflow of mixed documents: vendor packets, applications, claims files, contracts and shipping papers. Many already own an OCR or IDP tool. It gets the clean pages right and leaves an exception queue behind: low-confidence fields, unknown document types, several documents scanned into one PDF, pages from the wrong customer. That queue is the work most teams want to outsource.
To be clear about what we are: we do not sell document software, and we do not run OCR. We are the trained people who work your exception queue and check the machine's output, so the data you rely on has been looked at by a person where it matters. We also do not handle physical paper. If your archive is on paper, your team or a scanning provider scans it, and we take over from the scanned files.
What our team does
- Sort incoming files by document type, using your list
- Split multi-document PDFs and rename each file to your naming rule
- Index files by customer, account, date and document type in your document system
- Check OCR or IDP output against the page and correct low-confidence fields
- Key the fields the tool could not read
- Track expiry and renewal dates from certificates, licenses and contracts
- Route exceptions, such as missing pages or pages from another customer, to your team
For field-level verification of invoices, with an example of how OCR errors are logged, see document data extraction. For claims and policy files, see insurance document processing. If your sources are mostly simple forms rather than mixed documents, data entry outsourcing may be the closer fit.
How it runs
- You share the document types and rules. Send 20 to 50 sample files that show the real mix, your list of document types, your naming and indexing rules, and access to the tools.
- We agree scope and train. Our ops team sets the people and hours in your pilot plan. The team trains on your samples, and we build a gold set of files with known-correct types, splits and fields for you to confirm.
- We deliver, with a named team lead and QA sampling. The team works the live queue. The team lead handles unknown document types and rule changes, and sends open questions to your contact.
- You get a weekly report. Files and pages processed, QA sample results, exceptions by type, and the fields your tool misses most often.
Illustrative example: a vendor onboarding packet
Illustrative only. Names and files are invented.
- Input: scan_0917.pdf, 9 pages, emailed by a new vendor
- Your rules: one file per document · name as date_vendor_doctype · key the insurance expiry date into the vendor record
- Output, pages 1 to 3: vendor application, saved as 2026-09-17_NorthBay-Supply_application.pdf
- Output, pages 4 to 5: certificate of insurance, saved as 2026-09-17_NorthBay-Supply_COI.pdf · expiry 2027-03-31 keyed into the vendor record
- Output, pages 6 to 7: bank letter, saved and indexed · account number kept out of the file name, per your rule
- Output, page 8: blank page, removed and logged
- Output, page 9: a tax form for a different vendor (Rios Hardware), moved to the exceptions queue for your team
- Who did what: one team member split, named and indexed the packet. A QA reviewer re-checked the split and the file names as part of the weekly sample.
How we keep document work accurate
Document work goes wrong in two places: the wrong document type and the wrong field value. The gold set covers both. QA reviewers re-check a sample of processed files, comparing the split, type, file name and keyed fields with the source pages. They do not re-check every file. The weekly report shows the acceptance rate on the sample and lists the document types and fields that cause the most errors, so you can adjust your tool's templates upstream instead of paying people to fix the same miss each week. See how we check quality.
Pricing scope
Document processing in English is a standard task at $5 per person-hour, with a named team lead and QA sampling included. Work starts with a two-week prepaid pilot: people × hours per week × 2 weeks × $5, with at least 20 hours per person per week. After the pilot, we invoice weekly in arrears. Documents that need specialist knowledge, such as medical coding or legal review, documents in other languages, and setup of a new extraction tool are quoted after our ops team reviews the task. See the pricing page.
When to use the enterprise team instead
If the files you need processed are camera frames, drive recordings or other sensor logs that must be labeled for a perception model, that is not document work. The Deepen AI enterprise team scopes and prices it per project. See enterprise AI data services.
Send a sample of your real document mix and your rules. We email you a pilot plan with a quote.
FAQ
Are you an intelligent document processing company?
No. We do not sell document software. We are the trained people who work alongside it: we handle the exception queue, review low-confidence fields, key what the tool could not read, and log the errors it makes so your team can tune it.
Do you offer OCR services?
We check and correct OCR output. Your tool or provider runs the OCR, and our team compares the extracted fields with the page, fixes what is wrong, and records each error type in the weekly report.
Can you digitize paper documents?
We work from scanned or digital files. We do not handle physical paper. If your archive is still on paper, your team or a scanning provider scans it, and we sort, index, name and key the scanned files.
Which document systems do you work in?
Your document management system, shared drives, IDP review screens, accounting or claims systems, and spreadsheets, on accounts you create and can switch off at any time.
How do you handle personal data in documents?
Tell us in the pilot form. Before we start, we agree in writing who can see it, which fields they need, and what gets masked. We can limit that work to a named group within your team. See the security page.