Machine learning can provide many benefits to organizations and it is definitely not only applicable to large corporations. The trick is to translate the complexity into practical applications. With many technologies and tools available, the companies who apply them best will succeed in the market. This will require a combination of the technical expertise, the context and the ability of people to work effectively with outputs generated through machine learning and artificial intelligence.


Although many machine learning cases focus on numerical data, it can very well be applied to text-based data. During the last decade, there has been a strong development regarding text analytics. With UIMA from IBM the initial focus was on tokenizing and extracting information from the text such as phone number of name etc. Over the years many new techniques have developed and been added to the repertoire of text analytics. Sentiment analyses techniques, for example, is developed.

Depending on the nature of the question, machine learning can be applied as well. This can be as simple as clustering and classification to more in-depth analytics where alerts are generated in case references in a text are changed. Key will be that feedback loops can be established. An example could be Sharapova monitoring WADA regulations based on her own medical file. Medical support staff can provide the feedback loop to determine the relevance of alerts generated.

As an example to facilitate these discussions, KENTIVO has built a proto-type SPAM detector for SMS messages using Google TenderFlow.

Automatic Pattern Discovery

You know there are a lot of insights out there. They are in the heads of people as well as hidden in the information streams flowing through your company. To get testable ideas automatically could help you greatly. Might your processes need some optimization and improvements? Your product information in combination with trends in the market might hold little gems? With machine learning as part as our advise engine new opportunities are opened up to you.


Maintaining business rules is cumbersome and requires a lot of context and experience. This makes organizations highly depending on key people while at the same time, changes can be tricky. What if you leverage machine learning to automatically monitor and adapt business rule settings when there is a need? What if you can get warnings when your business rules no longer reflect the environment they operate on.


In the wonders of analytic and machine learning, the world is not perfect. Also, analytical models might have to change over time or when new types of insights become available. Simple things such as seasonal re-calibration or adapting to changes in data streams all come into play. Leveraging machine learning in conjunction with the operation of the organization requires proper management of the analytic models.