New report: The AI Ship Date Dilemma. 132 enterprise AI teams on the gap between a working demo and production.New report: The AI Ship Date Dilemma. Get the report

Skip to main content

Bespoke Invoice & Document Processing

Facilities and manufacturing · Accounts Payable: Invoice Processing

Thread AI Editorial

September 28, 2026

A document services provider gives each customer its own Lemma Worker, which reads supplier invoices as they arrive, learns from how past exceptions were resolved, and holds every posting for the customer's approver.

The Challenge

Most accounts payable work begins with a document: a PDF on an email, a scanned delivery note, or a packing slip that arrives with the goods. The providers that serve these teams have already captured this document, however, each invoice then passes through a long chain of checks:

  • Buyer and supplier identification
  • Whether the two actually trade with each other
  • Accounting parameters and dates
  • Duplicates
  • VAT and legal rules
  • PO matching

Each check is run once. If and when an invoice fails, it's flagged and routed to an operator, who makes a judgment call and moves on. When the same flag shows up on another invoice, someone has to make that call all over again. There’s no self improvement loop to catch the same error twice. Nothing learns, so every exception costs as much the hundredth time as the first, and some errors aren't flagged at all, so they go straight into the ledger as wrong output.

The Solution

The provider can build that memory into its service. Using Lemma, it deploys a Worker for a customer, assigns a unique identifier for each, and loads directly into the customer's own isolated tenant. The Worker also remembers how the customer’s operators have resolved past exceptions.

When an invoice fails a check, the Worker looks for a comparable case the customer's operators have already decided. If it finds one it's confident in, it applies that decision and moves the invoice on to the next check. If not, it sends the invoice to a person, and that person's decision is saved for the next case like it.

Over time, more exceptions resolve from precedent and fewer invoices reach a person. Nothing posts to the ERP until the customer's approver signs off.

Capabilities

1

Sorts every document before reading it.

Turns email attachments and device scans into one structured record, and sets aside anything that isn't an invoice before reading a single field.

2

Scores every field it extracts.

Gives each field a confidence score, measured against a rubric accessible by anyone reviewing the invoice.

3

Checks precedent before escalating.

Before a failed supplier match, purchase order match, or business-rule check reaches an operator, the Worker looks up how the customer resolved similar cases. Each operator decision is saved with its reason, and an invoice can come back for review at a later check in the same Run, carrying forward what was already decided.

4

Keeps each customer separate.

Each customer's Worker runs on its own isolated tenant, managed by the provider, with its own connection to their customer's ERP.

Tell us about your process.

Contact our team