Two-Way Matching

Two-way matching with your exceptions encoded.

We encode your match and exception rules — PO tolerances, no-PO logic, contract-rate enforcement — so only true exceptions reach a person.

  • AI implementation services
  • Two-way match before payment runs
  • Your exceptions, encoded
  • No new system for your team

Built with the leading AI platforms

NVIDIAAnthropic
Peter Enestrom, founder of Zaigo

Led by Peter Enestrom and the Zaigo AI & engineering team

YaleColumbia UniversityMicrosoft

The broken Friday

The bills keep coming. Nobody can say what we owe.

From AP inboxes we have sat inside: PE-backed services businesses running 2,000+ vendors, with every match and exception carried in someone’s head.

  • Bills with no PO behind them

    Invoices arrive with no purchase order to match against — two-way matching is impossible, so the bill waits while someone works out who approved it.

  • Burned by the extraction pitch

    The last tool promised automation and delivered 20% accuracy. Fields get extracted; a human still decides what to do with them. The bottleneck was never the reading.

  • The inbox at scale

    Three thousand invoices a day through one chaotic inbox — a 15,000-bill backlog, 130 people matching invoices by hand, and the match still loses to the mail.

  • Price creep nobody checks

    Nobody checks invoice prices against the contracted price sheet since she left. Hauler rates creep until the quarterly review, and pass-through charges get eaten instead of billed.

  • The 70-page supplier PDF

    One supplier sends a 70-page PDF someone splits by hand into 34 sub-invoices. Payment files fail, duplicate invoices slip through, and bills for work never done get paid.

What manual costs

Hand-matching is a headcount, not a workflow.

Reported operating examples show where manual work adds up. Your audit establishes the workload, error rates, and costs in your business.

20%Extraction accuracy of legacy OCR on emailed and mailed invoices
3,000Invoices a day through one chaotic AP inbox
130People hand-matching invoices at one AP organization

From anonymized engagements — waste services, facilities, distribution

How it works

We encode your exceptions.

No new system for your team to learn. The match rules your best AP person carries in their head become the workflow.

  1. 01

    Shadow the inbox

    One week inside your AP inbox catalogs every exception type: the bill with no PO, the stale PO, the rate that crept. That becomes your written rule set.

    Week 1
  2. 02

    Encode the match

    Rules get written against your vendors and contracts: PO tolerances, no-PO logic, contract-rate enforcement, approval routing. Matching invoices to purchase orders becomes automatic where the rules hold.

    Weeks 2–6
  3. 03

    Touchless, then exceptions

    Invoices that meet tested rules can advance to the agreed approval or posting step. Uncertain matches and rule failures go to a review queue, with controls defined by your finance team.

    Ongoing
Not another tool

Invoice matching software matches invoices. Your matching rules are the product.

Invoice matching software extracts fields, then hands the bill back to a human — the invoice with no PO, the stale PO, the rate that crept. The demo runs on clean PDFs; your inbox doesn’t.

Matching invoices to purchase orders is a rule problem, not a reading problem. AI agents do the reading; we encode your exception rules into the flow you already run — the two-way match happens before payment.

Finance teams describe difficulty checking high transaction volumes against contract rates, pass-through terms, and payer schedules. The workflow needs to account for those sources, with tested checks and a review path for missing or uncertain information.

Peter Enestrom, founder of Zaigo
Who builds it

Led by Peter Enestrom.

Co-Founder — leads AI & Engineering

Pete Enestrom

Every engagement is led personally by Pete, working with the Zaigo AI & engineering team from the two-week audit through the production handover. The person who scopes the work is the person who builds it.

Education
Yale & ColumbiaGraduate
Background
Microsoft & IntelFormer
Experience
Exited FounderVenture-Backed

Background

Questions

Asked by controllers and AP managers.

The straight answers, before you book anything.

What’s holding your business back?

A workflow ready for automation. An AI product you want to build. A problem that has sat on the roadmap for years. Let’s talk about what it would take to solve it.

Talk to Zaigo