Why AI needs Document Scanning

Document Scanning in action

An intellectual revolution is just starting, which will impact people who work with their minds just as much as the Industrial Revolution impacted people who worked with their hands.  And, like the Industrial Revolution, companies who embrace it will thrive, and those who are left behind will struggle to service.

The technology powering the revolution is, of course, AI in the form of Large Language Models.  ChatGPT and its rivals are already widely used, and remarkably good at what they do.  Professional companies like lawyers, consultants and accountants are embracing them.  They also know they cannot be better than their rivals if they are using the same publicly-available tools.  The way ahead is to build their own Large Language Model, and train it on data specific to their business and skillset.

Like any other computer system, the better the input the better the output.  And, for LLMs, that also means the bigger the input the better the output.  The more information it has to learn from, the better the results.  That’s why it’s important that it should have access to all of a company’s information even if it is not in a computer-readable format, like paper,  microfilm or microfiche.  A lot of companies still have a lot of information in these legacy formats.

The solution is, of course, scanning.  Scanning, plus OCR (Optical Character Recognition),  makes all the archives of information available.  This helps the AI to give answers which are more intelligent, and gives a way to be ahead of the competition.  The old archives have a new value, which can only be realised by scanning.