IBM lets (more) Watson roam on other clouds

In a bid to bring its AI tools to users of other cloud offerings, IBM has added some of its Watson services to IBM Cloud Private for Data.

The latter is a platform based on Kubernetes, which allows the software to run on a variety of public and private clouds, including AWS and Microsoft Azure. It has been around since autumn 2017 – so nothing too new – and already included Watson Studio and Watson Machine Learning. With the announcement however it now also has microservices for using Watson OpenScale and the Watson Assistant available.

While OpenScale is meant to manage AI instances, the Assistant service lets users build conversational interfaces. According to IBM those go beyond traditional chatbots, as they should be able to determine when, for example, human assistance is needed during a customer interaction or when asking a couple of more questions will do instead.

Those modules will become even more useful later this year, since the company is planning to also bring its Knowledge Studio and Natural Language Understanding services to IBM Cloud Private, which could help those looking into building a more complex system.


If you’re wondering about the why of the Cloud Private offering in general, it may be rooted in the multi-cloud strategies businesses tend to stick to nowadays as an IBM Cloud Blog suggests. With vendor lock-in being a hot topic and the options of disaster recovery or a quick cloud expansion if the private one won’t do, it’s a good step to stay relevant – and open up services for those that wouldn’t traditionally go for the big blue.

One of the things still standing between IBM and the wider adoption of its Watson offerings however is the lack of people knowing their way around machine learning technologies. While other companies tackle this by making their products easier to use without (or only little) prior knowledge, IBM invests in training programs, as General Manager Rob Thomas writes in a blog post announcing “Watson anywhere”.

A learning and certification program should teach developers in a number of online courses or two day workshops in one of the IBM Labs the basics of artificial intelligence and machine learning and introduce them to – probably their own – tools to get going. And since no training seems to be worth anything without a certificate to prove it, IBM is working on a certification programme at the moment to attest the skills needed “to implement enterprise use cases”.

- Advertisement -