From IBM Watson to watsonx: History
Last week, I ran into the “We are interested in Watson, but exactly is it?” situation. Someone at a customer knew that “Watson” was the “IBM AI thing”, but wanted to be sure and asked for a quick overview. Because I use this blog for my notes, why not write down this quick cheatsheet for my own assistance (pun intended)…
IBM recently celebrated 114 years of corporate history. It started research in what today is called Artificial Intelligence (AI) in the 1950s (see 70 years advancing AI). Considering the Timeline of Artificial Intelligence, it is only after some thousand years of the initial concepts, but still early in today’s “era of AI”. And it included some well-known AI showcases like the chess computer Deep Blue in the 1990s, IBM Watson playing Jeopardy in 2011, or Project Debater in 2018/2019.
After the appearance of IBM Watson on Jeopardy, IBM started to use the catchy name for branding its AI-related efforts and products using “Watson”. It included rebranding products like IBM Data Science Experience (DSX) into IBM Watson Studio. And you may have noticed that Watson Studio is now called IBM watsonx.ai Studio.
In 2023 at the Think conference, IBM announced watsonx as the successor to the Watson brand. As of today, both Watson- and watsonx-branded products are available, as can be seen in IBM product documentation overview under “W”:
IBM watsonx product portfolio
The IBM watsonx portfolio covers both the full stack and the entire workflow to support AI-based applications and in addition includes AI-powered assistants for specialized tasks. Clicking on “Portfolio” reveals the following:
- watsonx.ai is the end-to-end developer studio and includes the former Watson Studio
- watsonx.data allows to build the data foundation with a hybrid, open lakehouse approach.
- watsonx.governance provides end-to-end governance for AI deployment. It includes functionality to manage and monitor “AI” (including models and what makes up those models).
- watsonx Orchestrate (no, not watsonx.orchestrate 😉) is the product to coordinate AI agents, assistants and backend (system) tasks. It is an evolution of what could be done in early versions of the next:
- watsonx Assistant is the foundation for building virtual agents and make them available through various communication channels.
- watsonx Code Assistant, watsonx Code Assistant for i, watsonx Code Assistant for Red Hat Ansible Lightspeed, and watsonx Code Assistant for Z are products for coding or administrative tasks in specific environments.
Many of the above products are available “as-a-Service” (managed, in public cloud) or as “Software” for deployment in private clouds and data center. In addition, IBM has started with solutions that are built for and integrate with the core product and offer additional tools. Examples are watsonx.data intelligence and watsonx.data integration that help to reap benefits from the data managed in watsonx.data or to build the pipelines and data flows to create the data lakehouse.
Sometimes, watsonx is abbreviated to “wx” (like “wx.data”) - but not in official communication. So much for my cheatsheet full of links to background reading, product documentation, free trials, demos, and more.
If you have feedback, suggestions, or questions about this post, please reach out to me on Mastodon (@data_henrik@mastodon.social) or LinkedIn.