Bulk receipt scanning

One of these asks your client to do something three hundred times.

Per-receipt apps optimise for the person holding the receipt. Period batches optimise for the person closing the books. Both are real workflows and the honest question is not which is better, it is which one your clients will actually keep up. Here are both, answering the same five questions.

Receipt OCR for accountants

The question underneath it

Will they photograph every receipt, every month, all year?

Some clients do. Most do not, and the difference decides which of these two is the right tool rather than which one has more features.

  • If yes

    Per-receipt captureNot us

    Optimises for
    The person holding the receipt, at the moment they are holding it.
    Assumes
    The client photographs every receipt, every time, for twelve months.
    Review happens
    One decision per receipt, spread across the year, whenever they arrive.
    When the assumption fails
    The accountant inherits the gap at quarter end anyway: a shoebox, a scan, or a folder of phone photos.
    What you end up with
    A timeline of captured items to scroll and approve.

  • If not

    Period batchThis is us

    Optimises for
    The person closing the books, at the moment they are closing them.
    Assumes
    The pile arrives all at once, in whatever state it is in.
    Review happens
    One sitting per period, and only the flagged rows need eyes.
    When the assumption fails
    The client still has to send the pile. Nothing here makes that happen for you.
    What you end up with
    One workbook per supplier, sources embedded, entries ready to post.
1,200page cap per batch; year-end backlogs finish overnight

One of these depends on a client doing something three hundred times. The other depends on them doing it once. That is the whole difference, and it is not a feature.

Scan the pile once, drop it once.

  • Built for hundreds of pages.

    A typical 200-page batch processes in about 12 minutes, with pages handled in parallel and hard ones retried automatically. Bigger batches, the year-end shoebox, finish overnight.

  • Mixed documents, one drop.

    Receipts, supplier invoices, and bank statements in the same PDF, in any order. Eleven languages, mixed currencies, thermal paper and photographed pages all flow through the same pipeline.

  • Review concentrated, not scattered.

    Instead of approving items one by one all year, you review one period in one sitting, and only the yellow-flagged rows need eyes. The rest reconciles against printed totals.

  • Clients can still photograph.

    Phone photos are fine, they just go into the period PDF instead of a per-receipt app. No client training, no app adoption project, no discipline dependency.

  • Books, not a feed.

    The output is one Excel workbook per supplier with sources embedded, plus entries that post to your connected ledger. The deliverable is the period closed, not a timeline to scroll.

Bulk scanning questions, answered.

When expenses need capturing in real time by the person spending, field staff with mileage and meals, employee expense claims, businesses that genuinely keep the habit year-round. If that is your workflow and it works, keep it. Pileform is built for the complementary case: the period that arrives as a pile at close.

Yes, phone photos are normal input. The difference is where they go: into the period PDF alongside everything else, rather than into a per-receipt app at purchase time. Combine scans, photos, and email attachments into one PDF and drop it once.

A typical batch is 200 pages and processes in about 12 minutes. Larger batches, 500 to 1,000 pages, the classic year-end backlog, finish overnight. One batch can be up to 1,200 pages; you get an email when the workbook is ready.

No, the opposite. Extraction is per line, not per receipt: each line carries its description, amounts, and VAT rate, totals reconcile to each document's printed totals, and the original page is embedded next to its rows. Bulk refers to the input, not the granularity.

Rows come back in two states: high-confidence and ready, or yellow-flagged where something was inferred or uncertain. A period's review is typically about twenty minutes on the flagged rows, concentrated in one sitting, instead of item-by-item approvals scattered across the year.

Send the pile. All of it.

Sign up free, no card. Drop your messiest batch and compare the output to your current workflow before you commit to anything.

Start free