IBM today announced that it is freeing its Watson-branded AI services — like the Watson Assistant for building conversational interfaces and Watson OpenScale for managing the AI life cycle — from its own cloud and allowing enterprises to take its platform and running it in their own data centers. In a way, you can think of this as Watson as a managed service.
“Clients are really struggling with infusing AI into their applications because the data is distributed in multiple places,” IBM Watson’s CTO and chief architects Ruchir Puri told me when I asked him for IBM’s reasoning behind this move. “It’s in these hybrid environments, they’ve got multiple cloud implementations, they have data in their private cloud as well. They have been struggling because the providers of AI have been trying to lock them into a particular implementation that is not suitable to this hybrid cloud environment.”
So with this decision of bringing Watson to any cloud, IBM wants to give these businesses the option to bring AI to their data, which is significantly harder and costlier to move, after all. Puri also stressed that many enterprises have long wanted to use AI to make their operations more efficient, but they needed to run their AI tools in an environment they control and feel comfortable with.
At the core of the technical specifications for running Watson in their public or private cloud is IBM Cloud Private, the company’s private cloud platform that uses open-source technologies for running tools and services like Kubernetes and Cloud Foundry. That’s the platform that allows enterprises to then run Watson, too (which itself runs on containers, too).
Right now, the focus of this fire launch is on Watson Assistant and Watson OpenScale. “The capabilities we are releasing right now are based on our two flagship products. That addresses a very large domain of use cases that we come across,” said Puri. “In the remaining part of the year, we will bring the rest of the capabilities [to the platform]. For example, Watson Knowledge Studio will come along with it as well, as well as Watson’s natural language understanding capabilities that we currently have available in our public cloud environment will be ported on to it as well.”
With that, Puri argues, IBM will offer enterprises a full spectrum of tools for developing and running AI models using structured and unstructured data, as well as a full monitoring and life cycle management suite.
In addition to this, IBM also today announced that it is launching a new version of its Watson Machine Learning Accelerator that brings high-performance GPU clustering to Power Systems and X86 systems and which promises to accelerate AI performance up to 10x.
The company also today announced IBM Business Automation Intelligence with Watson, though it didn’t quite delve into the details. This new service, the company says, will give business leaders the ability “to apply AI directly to applications, strengthening the workforce, from clerical to knowledge workers, to intelligently automate work from the mundane to the complex.” I’m not really sure what that means, but I’m sure the business leaders who buy this service will figure it out.