International Business Machines (IBM +0.33%) is responsible for one of the most iconic moments in the history of artificial intelligence. Back in 2011, its Watson AI system won a three-game Jeopardy! tournament against two of the game show's top champions.
The tech giant tried to leverage that moment to turn Watson into a viable business -- but failed. Healthcare was a big area of focus, but that effort fell flat. Last year, IBM sold part of its Watson Health business to a private equity firm. The company still makes use of AI across its business, using the technology to help customers boost productivity and cut costs, but it quit trying to turn Watson into a household name.
NYSE: IBM
Key Data Points
Watson with a twist
Now, however, with the hype around artificial intelligence at a fever pitch following the launch of OpenAI's ChatGPT late last year, IBM is looking to take advantage. On Tuesday, the company unveiled a new AI platform called watsonx that will enable its enterprise customers to build and scale their own AI models.
IBM's watsonx is composed of three parts. The first part is an AI studio, set to be available in July, that allows customers to train and deploy machine-learning and generative AI models. The studio includes IBM's own library of AI models, as well as additional models and datasets developed through a collaboration with AI start-up Hugging Face.
The second part of the watsonx platform is a data store built on the Data Lakehouse architecture, which combines features and capabilities from data lakes and data warehouses. Data stores can be run on-premises or in multicloud environments, and are aimed at helping customers better leverage their data for analytics and AI tasks.
The third part of the watsonx platform is an AI governance toolkit, which will be available later this year. This will essentially help customers make sure their AI models don't go off the rails. Generative AI models can produce odd results, and maintaining the privacy of data is critical. The toolkit will help keep AI models in check.
IBM also plans to use the AI models available as part of watsonx to improve its own software products. For example, it will launch Watson Code Assistant, which will use generative AI to help developers produce code.
A focus on the enterprise
IBM's watsonx platform fits nicely with the company's general AI strategy: Use the technology to help its enterprise customers improve their own operations. This is the flashiest thing related to AI that IBM has talked about in quite a while, but ultimately, it's aimed at enterprises that want simple ways to leverage the power of AI.
On top of these software offerings, the watsonx platform cloud should also give a boost to IBM's sprawling consulting business. IBM is creating a "center of excellence" within its consulting organization, with more than 1,000 AI experts who will help customers adopt AI technology. Consulting is big business for IBM -- the segment brought in $5 billion in revenue in the first quarter, or about 35% of the company's total.
While AI products like ChatGPT make headlines, IBM's focus is more behind the scenes. As enterprises look to tap into the incredible potential of AI, the watsonx platform may represent the path of least resistance, particularly for existing IBM customers.
It has been more than a decade since the Watson brand emerged. Now, IBM may finally have a Watson product that could be a major success.
