Whereas until now engineers tackled a problem by programming, machine learning opens up a whole new set of possibilities. No more programming and testing an algorithm yourself. You define a learning model, feed it with ground truth data and little by little your machine learns to solve the problem itself.
To process documents, we work in the same way. Our own statistical learning model is called AUDI (Automatic Document Identification) . It learns to analyze the layout and knows how and where to find information in a document. Your input as a user serves here as truth data, which means that adjustments are made continuously. New document from a new supplier? No problem, it's just processed as well.
But we like to go even further than that. Language interpretation (NLP) allows us to convert free text into structured data or classification elements. An example? The mention 'discount cash' is used on invoices in many different ways. We use a cloud NLP model to convert all these variants into neatly structured information, which the system then cleverly passes on to the accounting department.
Machine learning models that have been trained for a specific domain create a whole host of new possibilities. The data that you can draw from all the documents in your company is therefore worth its weight in gold. It's called 'Data is the new oil', and that's not a lie. What's more, there's a lot going on in the industry. What used to be called 'non-automated' is gradually becoming so. At the same time, all kinds of tools and hardware are being developed that make your life and ours a little easier every day. Undoubtedly will be continued!