← All demosExecutive Lab · Automate

Document Intake Desk

Invoices, referrals, booking requests and service emails arrive all day, in every format. Watch one get read, checked against your systems and routed, then try the copy with a problem in it.

Sample documents from fictional companies. The demo reads only these samples, never your files.

Supplier invoice

Tallgrass Home Goods Co. Prairie Junction, Kansas

Invoice

Invoice no.: TG-48213
Date: October 6, 2026
PO no.: PO-77104

Bill to Cedar & Finch Home Stores, Accounts Payable

SKU	Item	Qty	Price	Amount
TG-1042	Linen throw pillow, sand	120	$14.50	$1,740.00
TG-2210	Stoneware mug set, 4 pieces	80	$18.75	$1,500.00
TG-3307	Cotton bath towel, slate	60	$15.70	$942.00
Total due $4,182.00

Terms: Net 30, due November 5, 2026

7 fields read, 5 checks run: handled automatically.

Extracted fields

Supplier Tallgrass Home Goods Co. 99% confidence
Invoice number TG-48213 99% confidence
Invoice date October 6, 2026 98% confidence
PO number PO-77104 97% confidence
Line items 3 lines, 260 units 96% confidence
Total due $4,182.00 99% confidence
Terms Net 30, due November 5, 2026 95% confidence

Checks

Passed

Purchase order PO-77104 is open and matches the supplier
Passed

Quantities match what the warehouse received

Receiving record RCV-30958: 120, 80 and 60 units.
Passed

Prices match the purchase order
Passed

Line items add up to the total
Passed

Not a copy of an invoice already paid

Handled automatically

Approved for payment

Every check passed, so it was approved without anyone touching it.

Bill created in the accounting system
Payment scheduled for the due date, November 5
Invoice filed with the purchase order and the receiving record

Structured data (JSON)

{
  "document_type": "supplier_invoice",
  "supplier": "Tallgrass Home Goods Co.",
  "invoice_number": "TG-48213",
  "invoice_date": "2026-10-06",
  "po_number": "PO-77104",
  "lines": [
    {
      "sku": "TG-1042",
      "description": "Linen throw pillow, sand",
      "quantity": 120,
      "unit_price": 14.5,
      "amount": 1740
    },
    {
      "sku": "TG-2210",
      "description": "Stoneware mug set, 4 pieces",
      "quantity": 80,
      "unit_price": 18.75,
      "amount": 1500
    },
    {
      "sku": "TG-3307",
      "description": "Cotton bath towel, slate",
      "quantity": 60,
      "unit_price": 15.7,
      "amount": 942
    }
  ],
  "total": 4182,
  "terms": {
    "terms": "net_30",
    "due_date": "2026-11-05"
  },
  "checks": [
    {
      "check": "po_open",
      "result": "pass"
    },
    {
      "check": "received",
      "result": "pass"
    },
    {
      "check": "prices",
      "result": "pass"
    },
    {
      "check": "math",
      "result": "pass"
    },
    {
      "check": "duplicate",
      "result": "pass"
    }
  ],
  "route": "approved_for_payment"
}

What this means for your business

Clean documents go straight through, so your team’s time goes to the exceptions instead of retyping.
Problems are caught before they cost money: a wrong total, a missing ID or a rate below the floor never slips into your systems.
Every document leaves structured data and a reason for each decision, ready for reporting and audits.

How it works

The document is read into named fields, each with a confidence score. Any field read with under 90% confidence goes to a person for a quick look.
Each field is then checked against the systems it touches: purchase orders, inventory, calendars, insurers or customer accounts.
If every check passes, the document is handled automatically. Otherwise it goes to the right person with the reason and the document attached.
The output is structured data that any system can use, shown at the bottom as JSON.

Want this working in your operations?

Start a conversation