Reading Financial Documents with AI A vendor-neutral guide for AI engineers argues that document ingestion pipelines in banking fail not on reading PDFs but on trusting the extracted numbers, citing the case where an AI system reads the accounting notation "(486)" — meaning negative 486 thousand euros of EBITDA — and stores 486, flipping the sign so a breached leverage covenant appears to pass. The guide, drawn from a year of building and comparing document ingestion pipelines in the financial sector, covers accounting notation, locale, scale and currency headers, comparative periods, layout, scans and spreadsheets, and lays out retrieval versus extraction, three architecture archetypes from RAG to per-page extraction with proof, and how to measure quality and pick an archetype. Every quarter, a corporate borrower sends its bank a PDF. Somewhere around page 12, there is a small table with a number in brackets: 486 . That is EBITDA, in thousands of euros, and it is negative. A credit analyst copies it into a spreadsheet, recomputes net debt to EBITDA, and checks the result against the covenant in the loan agreement. If the ratio breaches the threshold, the bank and the borrower are about to have a difficult conversation. Now imagine an AI system reads that PDF instead. It sees “ 486 ” and stores 486 . The sign has flipped. The leverage ratio looks healthy, the covenant passes, and nobody notices. Most document-AI demos show a chatbot answering questions about a PDF. That is nice, but it is not the problem banks have most of the time. The everyday problem is turning documents into useful data i.e., values that flow into credit models, risk limits, and regulatory reports, where a near miss is a defect . Multimodal LLMs made the reading part dramatically easier. They did not make the trusting part any easier. I have spent a good part of the last year building and comparing document ingestion pipelines in the financial sector. This post distills what I learned into a vendor-neutral guide for AI engineers. I’ll cover: 1. Why financial documents are hard for AI, and the specific traps that I’ve seen breaking real-world pipelines. 2. Retrieval vs. extraction: the single decision that shapes everything else. 3. Three architecture archetypes, from RAG to per-page extraction with proof, with Python code for the parts that matter. 4. How to measure quality and how to pick the right archetype for your use case. Why Financial Documents Are Hard A PDF is not a document in the way you as a human may think of it. It is a set of drawing instructions: put this glyph at these coordinates, draw a line from here to there, paint this image in that rectangle, etc. There is no “table”, no “row”, no “column header”. Those are things you see, because your brain groups the glyphs. Financial reports add their own traps on top of that: - Accounting notation. Negative values are printed in parentheses: 486 means −486. A dash can mean zero or “not applicable”. - Locale. 1 234 is one number with a space as thousands separator. 12,5 % uses a decimal comma. The same report may mix both conventions. - Scale and currency. A header says “EUR thousands” once, and every number on the page inherits it. Miss the header and you are off by a factor of 1,000. - Periods. Most tables show the current period next to a comparative one. Pick the wrong column and you get last year’s number, which is plausible and therefore dangerous. - Layout. Merged headers, multi-page tables, footnote markers glued to numbers 1 234¹ , and key figures that only appear inside a chart. - Scans. Signed compliance certificates and older annexes are often images with no text layer at all. - Slides and spreadsheets. Investor presentations put numbers in charts and text boxes with no reading order. Excel files are not documents at all: they are small programs with formulas, hidden sheets, and several tables per tab. None of this is exotic. It is what a normal quarter looks like for a credit, research, or KYC team. The Building Blocks Every ingestion pipeline I have seen is assembled from the same handful of components. What differs is how they are combined, and which one is trusted to have the last word. 1. Get the text. Text-layer parsers such as pdfplumber https://github.com/jsvine/pdfplumber and PyMuPDF