As the world goes digital, the majority of expense documents are still trapped inside paper receipts and invoices. Sadly, humans typing and inputting these documents is still, the most common form of expense data entry worldwide. Semi automated solutions using OCR and document templating whilst somewhat better, still leave data engineers with a brutal amount of unstructured data to parse and classify. Whether outsourced to data entry teams or handled in house, standard solutions are slow to develop and unscalable when parsing data out of random formats. To add to the challenge, anomalies caused by badly captured images can make even advanced OCR unreliable and parsing clean data virtually impossible. Inefficient and expensive data entry tasks need to be tackled seriously as part of any scalable and reliable expense solution. Enter EDE, (Expense Data Extraction) While many solutions focus on OCR using templating and text parsing to try and solve these never ending formats, we see far beyond this approach. We focus on intelligent EDE. Expense Data Extraction that truly understands formats and data fields. We don’t just see totals, we see languages, currencies, stores and businesses. We see times, dates, phone numbers and payment methods. We see product codes, barcodes, prices and quantities, and the list goes on. We also see anomalies too, solving crumples, folds, shading, bad light, and even warped text, enabling expense document data fields to be captured accurately, through a simple mobile phone photo. We read formats faster and more efficiently than a human, effectively delivering an intelligent, high performance EDE solution into any system in the world. We automate the reading of expense data, allowing data science teams to focus on greater value tasks and more creative solutions within companies. EDE technology frees companies from the burden of expense documents, empowering them to create innovative and winning products for their customers.