I’m Jose Antonio Parra, founder of Softnet Sistemas, an IT company in Alicante, Spain that has worked with small and mid-sized businesses since 1996. AIPROCESSIA grew out of what I kept seeing at those clients: a task as common as it is stubborn, which is typing supplier invoices into the accounting system by hand, one by one.

Here is an example, meant as an illustration and not as a figure from any study: in a small finance office, three hours a day spent keying in invoices adds up to about 60 hours a month (counting 20 working days) that don’t go to review, reconciliation or analysis. That math is why I started AIPROCESSIA, my project focused on intelligent process automation for SMBs.

I want to use this post to explain, in plain terms, what accounts payable (AP) and invoice automation actually look like once you strip away the hype, and what AIPROCESSIA’s invoice automation service (called Facturas IA in Spanish) does today. The gap between the marketing and the day-to-day reality is wide.

Updated October 20, 2026.

The real bottleneck isn’t the software. It’s the manual data entry

Keying in an invoice is slow, and it’s where mistakes creep in: a mistyped amount, VAT in the wrong field, an invoice entered twice. Modern AI-based capture reads a PDF or a photo of a receipt, extracts the data, checks that the numbers add up, and leaves everything ready to import into the accounting system, with no typing. The finance person goes from transcribing to reviewing only the invoices that don’t add up.

For a small business, the win isn’t just speed. It’s fewer errors, fewer duplicate payments, and a finance person who gets to do analysis instead of transcription.

What AIPROCESSIA’s invoice automation does today

Here is how the service works for supplier invoices. Today it targets Spanish accounting: it reads Spanish VAT and IRPF withholding (the income-tax withholding on invoices from freelancers and professionals) and works with Spanish charts of accounts.

• Invoices arrive on their own. By email to the company’s dedicated inbox (with authorized-sender control), or by uploading PDFs to the panel.
• AI reads the data. Supplier, tax ID, invoice number, dates, net amounts per VAT rate, withholding, totals and line items, without setting up a template for each supplier.
• It checks the math. Every invoice goes through an arithmetic check (net + VAT – withholding = total). If it doesn’t add up, it is flagged and doesn’t reach the books until someone reviews it.
• It suggests the account. The suggestion comes from your own chart of accounts (Excel or CSV), and you can correct it in an editor with the PDF alongside.
• It exports a batch ready to import. An Excel file in your accounting software’s format. It has been verified with GoldenSoft; for other software the format is set up during onboarding.

That is the whole circuit, not just OCR, and human review stays where it matters: on what doesn’t add up. Pricing is flat, with no per-invoice fees: from €49 a month + VAT for up to 100 invoices a month, and €99 a month + VAT for up to 300; guided onboarding (a one-time €150) is optional, and there is no lock-in. Prices and features are on the service page (checked on October 20, 2026).

Choosing a tool is harder than it should be

The category is crowded: Bill.com, Stampli, Tipalti, Quadient, AvidXchange, Yooz and a dozen others, and most comparison pages are written by the vendors themselves. When my team and I sat down to map the landscape honestly for the 2026 market, we ended up writing a full breakdown of the options, who each one is genuinely a good fit for, and where the pricing traps are. If you’re evaluating this for your own finance team, that comparison of AP automation software for 2026 is the most useful thing I can point you to. For how invoice OCR itself works and what to expect from it, there is also our guide to invoice OCR software in 2026.

Why I keep writing about this

I’m a firm believer that automation should be boring and reliable, not a science project. The agentic-AI wave makes a lot of promises; my job is to translate those into something a 12-person company can actually deploy and trust. That’s the lens behind everything I publish on AIPROCESSIA, and it’s the same lens I bring to the consulting work I do here.

If you run a small finance operation and you’re tired of the manual grind, start by measuring how many hours a week your team spends on invoice entry. That number is the clearest case you can build for deciding whether automation pays off for you.

Frequently asked questions

Is invoice automation the same as OCR?

No. OCR only reads the text in the image. Automation is the whole circuit: receiving, reading, validating, account coding, exporting, and human review where it’s needed.

What happens if the AI misreads an amount?

The arithmetic check catches amounts that don’t add up to each other and flags the invoice for review before it reaches the books. That is why human review remains part of the process.

How much does it cost to automate invoice entry?

AIPROCESSIA’s invoice automation costs €49 a month + VAT for up to 100 invoices a month and €99 a month + VAT for up to 300, with no per-invoice fees. Optional guided onboarding is a one-time €150. Details are at aiprocessia.com/en/invoice-automation.

About the author: Jose Antonio Parra, founder of Softnet Sistemas (Alicante, Spain, since 1996) and of AIPROCESSIA.