IBM has offered various forms of artificial intelligence for years, most famously its Jeopardy-winning Watson.
In its third-quarter earnings call on Wednesday, the company signaled that, like many other large tech companies including
"Nearly every business we talk to wants to leverage technology to offer better services, scale more quickly and fuel growth without increasing their footprint," said Arvind Krishna, CEO and chairman. "This has been driving demand for technologies that boost productivity and competitiveness like hybrid cloud and artificial intelligence."
Observers noted that generative AI using large language models is a relatively new area of emphasis for IBM.
"The company lags other providers on generative AI because those firms have been developing those products for several years," said Darrell West, senior fellow at the Brookings Institution. "It will take some effort to catch up with its competitors."
Krishna said he expects IBM's clients to use a combination of IBM's models, other companies' models, their own proprietary models and open-source models.
"This hybrid approach to AI is similar to the hybrid approach to cloud," he said.
IBM's generative AI software stack includes its Red Hat OpenShift software development platform. It also includes watsonx, a platform companies can use to train, tune, validate and deploy AI models, using prebuilt IBM models and open-source models hosted on Hugging Face.
IBM introduced some new AI technology in the third quarter, much of it based on watsonx.
In September, it announced the general availability of its first set of watsonx Granite generative AI models, which are based on watsonx and intended to let companies embed generative AI into their business applications and workflows. Earlier this year, IBM debuted watsonx Orchestrate, generative AI technology employees can use to automate tasks.
Use of such "digital worker" technology could bring companies productivity improvements of 4% to 10%, Krishna estimated.
In October, IBM launched watsonx Code Assistant, which developers can use to generate code.
"If you look at code, to a person, every one of the clients said we don't intend to get rid of any developers, but this allows us to take care of the tech debt that we've accumulated and makes every developer more productive," Krishna said. "Most enterprises are very quick to realize, if you are more productive, that means you have a competitive advantage to your peers. And if they have a competitive advantage to their peers, they'll take share without spending more on labor. That's an incredible long-term competitive advantage."
One of these code assistants, watsonx Code Assistant for Z, is designed to help clients modernize mainframe code and apps. It can translate COBOL code to Java, for instance.
"There are hundreds of billions of lines of code written in COBOL," Krishna said. "So the opportunity is significant."
Being able to convert COBOL to Java "is a big plus for organizations whose legacy IT systems date back several decades," West said.
This includes many banks with older core systems running on mainframes.
IBM also launched watsonx Assistant, a virtual assistant for customer service agents, in October. This could help companies save 20% to 30% in customer support costs, Krishna estimated.
"If you can have more calls, more chats answered by AI, that means you can have much more volume with a smaller number of people," he said. "Also, AI does not get tired. It doesn't get angry. It doesn't get upset. So there is an NPS improvement to get along with it."
Cloud giants are making major investments in generative AI and large language models. Their financial services clients are mostly in test-and-learn mode.
Microsoft, Google and Amazon also offer generative AI tools for developers, customer service people and workers in general. IBM may be the only Big Tech company with an assistant for translating COBOL to Java.
IBM has more than 20,000 data and AI consultants and a center of excellence for generative AI, Krishna said.
"Our book of business in the third quarter specifically related to generative AI was in the low hundreds of millions of dollars," Krishna said. "The interest is larger with thousands of hands-on interactions with our clients."
IBM's overall software revenue, including from its artificial intelligence technology, rose 8% in the third quarter. Consulting revenue rose 6% and infrastructure revenue fell 2%.
In a research note, Bernstein equity analysts said IBM's results were "not perfect, but just fine." They noted that the company "upheld its guidance" and that it has grown revenues almost 3% this year.
