Machines are smarter than me, at least after some smart people have told it what to think and since it is continually trained to get smarter. Maybe that computer really isn’t smarter than me, but it certainly has a better memory. In the context of our current lives, that translates into not forgetting passwords and never forgetting where I saved that file on the network.
Artificial intelligence and machine learning are all the rage in most industries, including the document management industry. There is a convergence of systems offering auto-classification and indexing of your files. Tagging the file with information that someone can use to find it later and dropping the file in some deep folder structure and forgetting where that was might go away. That seems like a smart approach and deserves some investigation.
There are a variety of ways to look at the auto-classification of files but the ultimate goals would be to make saving a file simple and finding a file reliable. There is no doubt that eliminating the need to tag or add meta data to a file before saving it is a time saver. Its also a time saver if the file, like an inbound invoice, could be identified as an invoice, auto start a workflow and automatically fill in the invoice number, date, vendor and amount. That’s very smart.
Clearly if users don’t need to remember where files are stored to reliably find them later, that is a time saver too. Think of all the times that you have spent clicking through endless folders to find some file or doing a keyword search, again and again. If the classification rules and machines are smart enough, you will never need to worry how cohorts are saving files. You will always find the information you need.
There are a few things to consider before adopting this type of solution. First, it will not work as well with documents that are being added to the repository via scanning. The OCR engines are still not 100% accurate and it can cause some hiccups. Second, cost may become an issue for some companies. Until this is a common feature across the industry, it will be costly to implement. Third, the technology works best if used on consistent data sets. For instance, the AI engine knows that all inbound documents are invoices and therefore can be highly tuned.
This is a very exciting solution for the document management industry. Over time it will become common place. In the meantime, it could work very well in certain circumstances and should be explored. Of course, thorough testing is needed on your specific data and use case. In the meantime, continue to look for efficiencies in your current approach, including workflows, smart tagging and user training. These are also very smart, and in most cases affordable solutions.
Millennia Group has been providing workflow and document management solutions since 1996. For more information visit www.mgdocs.com or send an email to firstname.lastname@example.org.