Autonomous vehicles will make driving simple. Get in, tell the car where to go, sit back and relax. Artificial intelligence is going to read our minds and do our work for us. The same goes for robots. High speed data via 5G wireless will transform our lives by making these and other advancements possible. There are hundreds of steps and parts just to make a simple pencil.
The fact of the matter is that behind the scenes lots of hard work is happening to make something seem simple for the end users. For example, what if artificial intelligence changes its mind or futuristic fabrics confuse the autonomous vehicle system so it doesn’t recognize a person in the street? To keep it simple requires ongoing work by the engineers and tech teams.
It isn’t really that AI will change its mind so much as the data that is fed into the engine may pose a completely different point of view. Just like the autonomous vehicle system will need to be retrained as our physical environment changes, so too will the AI system. These are just some of the issues that will require ongoing effort to keep the appearance of simplicity for the end user. After all, if its not simple, then it likely won’t survive as a solution.
As an example of a situation that will require re-training of the AI system, consider a corporate merger. Company A has always classified its legal contracts in three ways, short term, long term and specialty. Based on the contents of those three contract templates the AI engine can auto-classify them. However, Company B uses different classification schemes; standard, custom and master. After the merger, someone will need to reclassify the old documents so the system can re-learn.
It’s also possible that normal business transformation will take place altering document classification schemes. As new applications are created and sources of information expand, the knowledge base of the AI engine will need to be modified. This is evident already today with the newest document storage silos that hold corporate information, such as in chat bots and instant messaging systems. Soon documents will be in a blockchain database or replaced altogether by data. AI engines will not only need to get smart but will need to have access to all these systems.
As we plan for the introduction of AI into document management systems there are several things to focus on. First, it has to make the whole process of document management simpler for the user, that is goal number 1. Second, it also needs to have flexible architecture given the inevitable environmental changes that will occur. Lastly, plan for regular audits of the data. It will take an ongoing commitment to keep the information accurate – finding the right information easily and quickly is the business goal.