Convert purchase order PDF to Excel on Mac (2026 workflow)

Eleven purchase orders in the inbox by 9 a.m., and every one of them is a PDF. One came out of SAP, two out of NetSuite, one from Coupa, and the rest from whatever the customer’s ops team happens to run — each with its own layout, its own item codes, its own way of splitting a description across three lines. Your system needs the same thing from all of them: which items, how many, at what price, shipping where, by when. And the way that data gets in today is someone on the order desk retyping it, line by line.

Retyped POs are where fulfillment errors are born. A transposed quantity ships 120 units instead of 12; a mis-keyed item code picks the wrong SKU; a missed requested-ship-date turns into an expedite fee. The data was structured in the customer’s ERP — their system generated the PDF from a table — and the PDF is just the print layout standing between their table and yours.

This guide is the Mac workflow: convert a purchase order PDF to Excel on Mac — read on-device, never uploaded — with the PO number, item codes, quantities, and prices in their own columns and the header fields repeated on every row, ready for import or reconciliation.

Why purchase orders defeat table extractors

A PO looks simple — a header block and a line-item table. Four structural habits make it a document that grid-based extractors consistently mangle:

1. The data you need is split between the header and the table. PO number, customer name, order date, ship-to, requested delivery date live in the header block; the items live in the table below. Your import wants both on every row. A grid extractor sees two unrelated regions and hands you the table with its identity amputated.
2. Every customer’s layout is different. Template-based tools want you to map fields per layout — workable for one big customer, hopeless for the long tail. The eleventh PO of the morning is always from someone new.
3. Line items sprawl. Descriptions wrap across lines, and many POs print two item codes per line — the customer’s internal code and your catalog number. Grid tools split wrapped rows into fragments and merge the two codes into one cell, and both errors surface later as shipment mistakes.
4. The table ends in noise. Subtotal, freight, tax, and total rows sit inside the same visual grid as the items; below them, a wall of terms and conditions. Left in the extraction, they become phantom line items that inflate every quantity sum downstream.

Fixing these requires reading the document — knowing that “Req. Ship” in the header belongs on each row and that the “Freight” row isn’t an item. That’s the gap a prompt-driven extractor closes.

Method 1: ignitai on Mac (the read-on-device way)

ignitai extracts by description, not by grid coordinates, which is what makes one workflow cover every customer’s layout:

1. Drag the PDF in — or pick it with the file picker, or right-click → Open With from Finder. A scanned or faxed-and-scanned PO works the same way as a born-digital one; the scanned-PDF walkthrough covers scan-quality details.

2. Describe the output shape in plain English. A prompt is required — the Convert button stays disabled without one — and each structural problem above dissolves into a clause:

   “One row per line item: po_number, customer, order_date, requested_ship_date, customer_item_code, our_item_code, description, quantity, unit_price, line_total. Take po_number, customer, and the dates from the header and repeat them on every row. Skip subtotal, freight, tax, and total rows. ISO dates, bare amounts.”

   Save it as a preset — it syncs to your other devices via iCloud — and tomorrow’s inbox is one click.

3. Pick Excel, hit Convert. The PDF is read on your Mac — never uploaded — and only the recognized text goes to ignitai’s private hosted pipeline to build the sheet. On macOS 26, Apple’s native on-device table recognition makes dense line-item grids extract noticeably sharper; the same flow works back to macOS 14.4. A typical PO is usually ready in under a minute.

4. Open the result in Excel or Numbers. One worksheet, one row per line item, every cell clean text — ISO dates, bare quantities and amounts — so one formatting pass types the columns. The in-app preview is read-only (rename only); edits happen in the spreadsheet.

The day’s orders as one batch

The eleven-PO morning is the case batch mode exists for: select all the PDFs in Finder and drag them in together with the same prompt. They consolidate into one worksheet — one row per line item across every order, ready to sort by requested ship date or group by customer.

Three things keep the batch honest:

• There is no filename column. Filenames never leave your Mac, so the sheet can’t say which file a row came from. The identity has to come from the documents themselves — which is why the prompt puts po_number and customer from each header on every row. On a purchase order that’s no hardship; the PO number is the most prominent field on the page.
• Images run separately. If one “PO” is actually a photo of a printed order, a drop that mixes PDFs and images imports only the PDFs — run the image as its own batch with the same prompt.
• A failed file stops the job. There’s no partial output or failed-file list; if one PDF is corrupt or unreadable, fix or drop that file and run the batch again. With a morning’s worth of POs that’s a ten-second retry, not a project.

Method 2: Excel for Mac’s Get Data (no PDF connector)

Excel for Mac ships Power Query under Data → Get Data, but the Mac connector list — Text/CSV, XLSX, XML, JSON, SharePoint, SQL Server, OData — has no From PDF option; that connector is Excel-for-Windows-only. Even on Windows, From PDF detects the visual grid per page: the header block arrives as a disconnected mini-table (or not at all), wrapped descriptions split into fragment rows, and the freight and total rows land inside your data. The repair work after Power Query is the data entry you were trying to eliminate.

Method 3: Preview + Numbers (the one-off fallback)

For a single short, born-digital PO: open it in Preview, select the table, copy, paste into Numbers, and repair by hand — then retype the header fields into new columns yourself. It’s tolerable exactly once. Multi-line descriptions paste as extra rows, and Numbers has no Text to Columns command, so every merged cell is a manual fix. For anything recurring, this is the workflow you’re escaping, not the solution.

Method 4: web converters (the confidentiality line)

Generic web PDF-to-Excel tools have the same per-page grid blindness as Power Query, and their free tiers gate at a few files — an inbox of POs exceeds that immediately. But the sharper problem is what a purchase order is: a customer’s negotiated pricing, their order volumes, their ship-to addresses, sometimes their entire product mix with you. That’s commercially sensitive data about someone else’s business, often covered by the confidentiality clause in your supply agreement. Uploading it to a converter you don’t control is a decision you shouldn’t be making on their behalf. The on-device read exists for exactly this class of document: the PDF stays on your Mac, and only recognized text touches ignitai’s private pipeline.

On iPhone and iPad

Order intake doesn’t always happen at a desk:

• iPhone — a printed PO handed over at a trade counter or job site: capture it with the in-app camera scan, apply the saved preset, and the line items are in a sheet before you’re back at the warehouse. An emailed photo of a PO goes in through the share sheet the same way.
• iPad — triaging the inbox away from the desk: save the attachment to Files, select it from the picker or share sheet (there’s no external drag-and-drop into the app on iPad), convert with the synced preset, and iCloud puts the XLSX on your Mac for the import.

One subscription covers iPhone, iPad, Mac, and Vision Pro, so the preset you wrote at the desk works at the counter.

After the sheet: import and the three-way match

The sheet isn’t the deliverable — the downstream steps are:

1. Order entry import. One row per line with consistent columns is the shape ERP and order-management imports want. Map our_item_code to your SKU field and the quantity errors that come from retyping simply stop happening.
2. The three-way match. With POs extracted to rows, matching what was ordered against what was invoiced becomes an XLOOKUP on PO number + item code instead of a side-by-side eyeball. The invoice-PDF-to-Excel guide covers the other half of that pairing — and for the delivery leg, what actually shipped, the bill-of-lading workflow turns the signed freight copies into the third set of rows.
3. Demand visibility. A quarter of extracted POs is a demand dataset: pivot by customer, by item, by month. If the destination is a shared tracker, the same XLSX imports straight into Google Sheets — the PDF-to-Google-Sheets Mac guide covers that handoff.

For the recurring high-volume case — hundreds of documents, not eleven — the batch-conversion guide goes deeper on consolidation prompts.

Bottom line

For customer purchase orders that need to become working Excel rows on a Mac: drag the PDFs into ignitai, write the one-paragraph prompt (line-item columns, header fields repeated on every row, skip the totals), pick Excel, Convert — one order or the whole morning’s batch. Use Preview-and-paste for a one-off; skip web converters entirely for documents that carry a customer’s pricing and addresses. The order desk’s time belongs on exceptions, not retyping.

Get ignitai on the App Store — sign in with Apple, first conversion free, then $19.99/mo after a 3-day trial; one subscription covers iPhone, iPad, Mac, and Vision Pro.

FAQ

How do I get purchase order line items into Excel on a Mac?

Drag the PO PDF into ignitai, describe the columns you want in plain English — PO number, item code, description, quantity, unit price — and ask for the header fields repeated on every row. Pick Excel and hit Convert. The line items come back as one sheet, one row per line, with the PO-level fields filled down.

Can I combine a whole day’s purchase orders into one sheet?

Yes — select all the PDFs and run them as one batch with the same prompt. They consolidate into a single worksheet, one row per line item. Ask for the PO number and customer name from each document’s header on every row; filenames never leave your Mac, so the documents themselves have to carry the identity.

Are the purchase orders uploaded?

No — each PDF is read on your Mac and never uploaded. Only the text recognized on-device goes to ignitai’s private hosted pipeline to build the sheet. A customer’s PO carries their pricing, volumes, and ship-to addresses, which is exactly the kind of document you don’t paste into a web converter.