Robots are smart, right? Well, maybe we shouldn’t tell them, they might read this, but reality shows that most of the robots are quite dumb. They can only work with structured data, and only in a way that we have learned them do.
So how do we solve this, lets add Artificial Intelligence to the robots! Or wait, maybe that’s a bit overkill, what is that thing ‘AI’, is it the humanoid robot that IT ordered to replace everyone? Probably not (yet), for enterprises we see a few large categories appear where this technology is applied
Natural Language Processing
One things robots are not so good at is interpreting free text, whether it’s an email, a contract or even spoken language, there is a lot to understanding the meaning, the semantics and understanding the context. Via complex analytics, this analysis is getting faster and faster, and soon will be near-real time.
Administration is both the enemy as a must have for all companies, but all that paperwork also can cause a headache, think about invoices, purchase orders, contracts, bills of lading, mortgages and many other documents with varying formatting and content. You don’t want to need to tell a system for each and every document exactly where to find it, Content Intelligence tools can Digitize, Classify and Extract data using deep learning algorithms, and neural networks to imitate how humans read the documents. Selecting the right tool here can save a lot of work.
Probably the most well-known ‘bot’, can be used both internal-facing and customer-facing, Often used as interface to unattended automations, or even directly to other IT systems, can be a great way of structuring data, making it ready for processing. Though a word of caution, also in the world of chatbots there are those who just finished primary school, and those who graduated from University. Look at the two Facebook chatbots that started inventing their own language, just because English wasn’t efficient, forcing the researchers to stop the project because no one understood what they were saying anymore.
Machine or Deep Learning
Data is power, those companies that have control over their data are the companies that will be successful. Understanding the customer and their needs, as well as creating a feedback loop on campaigns and the customer experience. Machine and Deep Learning are not a solution in itself but more a toolset that can be used, one of the important cores is that most solutions here have a dataset that you want to understand and will revolve around learning, But very tailored to a specific problem, like how you can win with chess or go.