Tax Document Automation

Automate tax document intake

Unstract sorts the W-2s, 1099s, K-1s and 1098s each client sends, checks them against last year's return and this year's tax organizer, and imports the figures into your tax software.

client-4417 · Okafor household, 2025running
Client docs11 documents
Sort & extract164 fields from 11 docs
Check vs prior1 K-1 missing
To tax softwareUltraTax CS
09:02:11Portal upload OKAFOR-2025.pdf · email 1099 · 4 scanned receipts · organizer
09:02:19Sorted · 11 documents filed to client 4417, 2025 return
09:02:31Read · 164 fields · lowest confidence 0.91 (receipt)
09:02:33Check vs 2024 return · Ridgeline Partners K-1 not received · held
09:14:50Preparer marked K-1 due · missing list sent to the client
09:14:52Imported to UltraTax CS · workpapers indexed · return ready to prepare

Trusted by tax and accounting teams around the globe

G2 Leader, Fall 2026G2 Best Usability, Summer 2026
Moody'sEndpointMedaxionPinpointConvrSmart ClerkCyberSoftThoughtFocusZluriIOA
The problem

Every client sends their season in a different shape

One client uploads a single scan of every form. Another emails a brokerage statement in March and a K-1 in April. Staff open each file, rename it, work out which return it belongs to and key the boxes, while the filing deadline gets closer.

CONSOLIDATED 1099MARAPR 15 DEADLINE

A 40-page brokerage 1099 keyed box by box in March

FORM W-2RECEIVED MAR 03

A W-2 photographed at an angle, with the wages box half in shadow

2024 RETURNK-1s2THIS YEARK-1s1

A K-1 that came last year and has not come this year

CLIENT INBOX · OKAFORFEBW-2 ×2FEB1098MARbrokerageAPR⚑ K-1

Forms trickling in by email from February to April

CLIENT 4417 · 1099s1099-INT1st Ohio412KEYED1099-INT1st Ohio412⚑ 2ND

The same 1099 uploaded twice and keyed into the return twice

Until every document is sorted, read and checked, a preparer can't start the return, and the gaps show up in April.

Unstract files each document to the right client and return, reads every box, checks the set against last year's return and the organizer, and imports the figures. Preparers start from a complete file or a short list of what's missing.

Book a consultationBring last season's client folders; we run a few of them live.
How it works

From the client's upload to your tax software, without the keying

Five steps, run on their own for every client. Parsing by LLMWhisperer, extraction by Document Studio, the whole run by Flow.

1 · Client document intake

Pick up the client's documents wherever they land

Clients upload W-2s, 1099s, K-1s, 1098s, brokerage statements and receipts to the portal, email them to the preparer, or drop off paper that the office scans. Each file is picked up where it arrives and tied to the client and the tax year, one record per return.

Nothing waits in a shared inbox for someone to save it to the right folder.

CLIENT · PORTALOKAFOR-2025.pdfCLIENT · EMAILBrokerage 1099OFFICE · SCAN4 receiptsCLIENT · PORTALTax organizerCLIENT 4417 · 2025forms6receipts4organizer1
2 · Sort and index

Split one upload into the forms inside it

Many clients scan every form into one PDF. The file is split at each form, and every part is classified by form type and tax year, then filed to the right client and return with a name your staff can find.

A 1098 scanned behind a K-1 is still read as a 1098, and last year's stray W-2 is filed to last year.

OKAFOR-2025.PDF9 PAGES · ONE UPLOADW-2 ×2pages 1–2✓1099-INTpage 3✓K-1 · 1065pages 4–7✓1098pages 8–9✓
3 · Read and extract

Read every form, then pull the fields that matter

LLMWhisperer reads each form with its layout kept, so a phone photo of a W-2 or a faint copy of a 1098 comes out as cleanly as a payer's PDF. Document Studio then extracts the payer, the recipient, the TIN, every box amount and the state details, on every form type the client sent.

The same field definitions hold whether a W-2 comes from a payroll provider or a small employer's own template.

FORM W-2employerBrightwater Co✓employeeMaya Okafor✓ssn***-**-6208✓box1_wages$84,250✓box2_fed_wh$9,870✓box17_stateOH $3,105✓confidence0.97box 12 codes
4 · Check against last year

Check it against last year's return before anyone touches it

What arrived is compared with last year's return and this year's tax organizer, payer by payer: a 1099 that came last year and not this year, a missing K-1, a consolidated statement whose totals don't add up, a name or TIN that differs from the client's record. A complete set keeps moving. A gap stops in a review screen with the expected form, where it came from and what arrived.

Expected forms, corrected forms, totals that disagree and low-confidence reads are the conditions you set; the rest of the client's documents carry on.

CLIENT 4417 · HELD⚑ K-1 MISSINGRECEIVED2024 RETURNORGANIZER1 K-1Cedar Fork LLC only2 K-1sSchedule E · page 22 K-1s✓client ticked yesMISSINGW-2s1099sRidgeline Ptrs LP2 of 2 received2 of 2 receivedASK CLIENTMARK NOT NEEDEDOpen the page →
5 · Into the tax software

Import the return data where your preparers already work

The extracted figures flow into your tax software, such as UltraTax CS or Lacerte, and the documents are indexed in the workpapers. Your practice management system, TaxDome or Drake Workflow, gets the status, and the client receives one list of what's still missing instead of three separate emails.

Preparers open a return with the data in place, and every decision on a missing form is recorded against the client.

CLIENT 441711 docs · 164 fields1 K-1 still dueTAX SOFTWARE✓ importedUltraTax CSWORKPAPERS✓ indexed2025 binderTAXDOME✓ list sentMissing: 1 K-1

Watch last season's client folders go through it.

Bring a handful of client folders, the messy uploads included. We run them live and show every form filed, every box read and every gap flagged against the prior-year return.

Supported documents & fields

Everything a client sends for the return

Documents

Fields captured

Form type and tax yearPayer namePayer TIN or EINRecipient nameRecipient SSN or TINClient and spouse namesAccount numberCorrected checkboxFiling statusDependentsAddress on the formPartnership EIN
Connectors

Our agents are extroverts.

Cloud storage like S3, Google Drive, Dropbox and Box. Databases and warehouses like Snowflake, BigQuery, Postgres and SQL Server. Accounting, ERP and CRM systems like QuickBooks, Xero, SAP, NetSuite, Salesforce and HubSpot. Project trackers like Jira, Asana, ClickUp and Linear. And anything with an API.

100+
Connectors ready to go right out the box
Amazon S3
Google Drive
Dropbox
Box
Microsoft Azure
Google Cloud Storage
Snowflake
BigQuery
PostgreSQL
SQL Server
MySQL
Amazon Redshift
Oracle Database
SAP
QuickBooks
Xero
Sage
Odoo
HubSpot
Zoho
Jira
Asana
ClickUp
Linear
Getting live

Running on a few clients before the season starts

No developers on your side and no change to how clients send their documents. The intake run is built on last season's client folders and proven on a group of returns before the whole client list goes through it.

01

Book a consultation

We go through last season's intake with you: how clients send documents, how your staff name and file them, and what a preparer checks before starting a return.

02

Built on last season's client folders

The form types, the prior-year checks, the tax organizer questions and the import into your tax software are set up on your own clients' documents, not samples.

03

Pilot on one preparer's clients, then widen it

One preparer's clients run live first. When the counts and the missing-document lists hold, the rest of the firm moves over, and the checks carry with it.

Who it's for

Built for the firms that prepare the returns

CPA and tax preparation firms

Every client's documents sorted, read and checked against last year's return before a preparer opens the file, so busy season starts with complete returns and a short list of chases.

Enrolled agents and tax resolution firms

Years of client records and prior returns filed by form and tax year, so a resolution case starts with the income history already read.

Bookkeeping and outsourced accounting firms

Clients' year-end forms collected in the same place as the books, with 1099s and 1098s read and matched to the ledger before the return.

Corporate tax departments

K-1s from investee partnerships and 1099s issued to the company read on arrival and checked against last year's provision workpapers.

Outcomes

Client documents sorted in minutes, gaps caught before April

<3 minfrom a client's upload to the documents read and in your tax software
91%of client documents filed and read without a preparer keying them
54%fewer hours spent sorting and keying client documents
2×faster to a complete document set for each return
Rather than spending the time manually sifting through unstructured documents, the system does the categorization of documents as well as extracting important information in a highly accurate manner.
G2Bogdan GrishinLead Full Stack Node JS Engineer · Reviewed on G2
1 / 3
Security

Client tax documents, handled with care

A client's tax documents carry Social Security numbers for everyone in the household, wages, bank and brokerage account numbers and the mortgage on the house: enough for someone to file a false return in the client's name.

AICPA SOC 2ISO 27001GDPRHIPAA
  • Encrypted in transit and at rest.
  • Client tax documents are never used to train models.
  • Access by role, client and engagement, with every review and correction recorded against the return.
  • Retention you set: processed documents are deleted on the schedule your firm's record-keeping policy sets.
  • Self-hosted deployment for firms whose security plan keeps taxpayer data inside their own environment.
See how taxpayer data is handled in the Trust Center
FAQs

What tax firms ask before busy season

Unstract classifies each document by form type and tax year as it arrives: W-2, each 1099 type, K-1, 1098, brokerage statement, receipt or organizer. It splits a single scanned upload at each form, files every part to the right client and return, and then reads the boxes.

Decide the folder and naming rules your firm already uses, then let the run apply them: Unstract sorts each document by client, return, form and year, indexes it in the workpapers and records what arrived against what was expected, so nobody renames files by hand.

Look for software that reads the boxes as well as filing the form. Unstract reads the payer, recipient, TIN, every box amount and the state details, checks them against the prior-year return, and imports the figures into your tax software, so staff review values instead of keying them.

Your tax workflow management software, such as TaxDome or Drake Workflow, tracks each return and talks to the client. Tax workflow software moves the return along; it doesn't read the forms. Unstract does the reading and checking inside that tax workflow: it updates the status, sends the missing-document request through it, and leaves the engagement where your team already manages it.

Yes. The answers in this year's tax organizer and the forms on last year's return set the list of expected documents for each client. A new rental property ticked on the organizer adds its documents to the list; a closed account drops off it.

The client's file stops in a review screen with the missing form, the payer and where the expectation came from. A preparer marks it due or not needed, and the client gets one list of everything still outstanding. The documents that did arrive are read and imported in the meantime.

A form with the corrected box marked is held beside the original so a preparer sees which figures changed before the return uses them. On a consolidated statement the 1099-INT, 1099-DIV and 1099-B sections are read separately, and totals that don't add up to the summary page are flagged.

Anything they have: payer PDFs, phone photos of a W-2, a single scan of every form, faint copies of a 1098, brokerage statements of forty pages, handwritten receipts and donation letters. LLMWhisperer reads each one with its layout kept, and low-confidence values stop for a person instead of reaching the return.

Your tax software, such as UltraTax CS or Lacerte, receives the extracted figures, and the documents are indexed in your workpapers. Flow posts them as part of the run, alongside the status update in your practice management system.

Unstract is SOC 2, ISO 27001, GDPR and HIPAA compliant. Client documents are encrypted in transit and at rest, never used to train models, and kept only as long as your firm's policy allows. Self-hosted deployment keeps taxpayer data inside your own environment; see the Trust Center.

We set it up with you in a consultation, on last season's client folders, your filing rules and your tax software, with no developers needed on your side. One preparer's clients run first, often within days, and the rest of the firm follows before the first W-2s arrive.

Next step

Bring last season's client folders to the consultation

A few clients' uploads, the single-scan PDFs and the phone photos included. You watch each form get filed and read, and every gap against last year's return shows why it was flagged.