your123 - stock.adobe.com
IBM targets agentic AI orchestration
The long-established vendor launched agent-building tools, prebuilt agents, and orchestration and integration capabilities for enterprises to integrate with major vendor platforms.
IBM on Tuesday introduced what it said are new technologies that help enterprises build and deploy AI agents. IBM's agentic tool suite adds to a market ripe with agentic platforms from cloud providers like Google, Microsoft and Salesforce.
The vendor, one of the most enduring tech companies due to its dominance in mainframe computing in the 1960s and 1970s, revealed new agentic capabilities in Watsonx orchestrate at its Think 2025 conference in Boston.
Watsonx orchestrate is a platform that lets users build and manage AI assistants and the autonomous and semi-autonomous agents that have swept the AI landscape as generative AI technology has quickly matured since it exploded onto the scene in late 2022.
IBM said Watsonx orchestrate includes tools that let enterprises quickly build agents in less than five minutes on any framework. The platform also has prebuilt domain agents specializing in areas such as HR, sales, and procurement. It is also integrated with more than 80 enterprise applications from providers including AWS, Adobe, Oracle, Salesforce, ServiceNow and Workday.
The platform also provides and agentic orchestration capabilities that can coordinate multiple agents and tools -- a concern as agents proliferate in enterprise IT systems. IBM also introduced a new Agent Catalog in Watson orchestrate that simplifies access to 150-plus agents and prebuilt tools from both IBM and its partners.
"Our view is there was a big gap in the market for how … you integrate agents. So it's not just about us providing agents, which we will do," said IBM chief commercial officer Rob Thomas, during a media briefing." "It's also enabling a company that's going to be looking at a future where there's hundreds, thousands, maybe millions of agents running at once."
A long time coming
According to Mark Beccue, an analyst with Enterprise Strategy Group, now a part of Omdia, IBM has been working on the new capabilities and tools in Watsonx orchestrate for a long time.
"They've been very thoughtful and forward-thinking about what agents can do," Beccue said. He noted that IBM presented agentic AI and coordination issues at the Think 2024 conference, before agentic AI became widely popular over the past year.
While agentic orchestration is still in its fairly early stages, IBM stands out in the agentic AI market by providing opportunities for enterprises to integrate with leading enterprise applications and tools for orchestration and observability.
"They're going to be a force to be reckoned with in this," Beccue continued. Moreover, with its focus on linking to major enterprise applications, IBM is differentiating itself from the new Google Agent2Agent Protocol because, instead of a standard, it is an integration, and IBM possesses experience helping enterprises integrate software.
"IBM is better at holding hands," he said. "Their whole job is system integration."
Other than capabilities for AI agents, IBM also introduced Watsonx.data integration, a single interface for orchestrating data across formats and pipelines. It also introduced Watsonx.data intelligence, a tool that uses AI technology to derive insights from unstructured data.
"Are they in a good position to help people who have data and stuff in a lot of different places?” Beccue said. “Yes, probably because they know that stuff and they're used to dealing with it, and so it sets them up for that.”
Early stage
However, despite its strengths, IBM is still in the infancy of what this technology can do, said Bradley Shimmin, an analyst with Futurum Group.
"Most of the vendors, IBM included, are chasing after the fundamentals that will go into building agentic systems that are manageable, governable, and securable," Shimmin said. He added that it’s promising for IBM to link the data integration and agentic product capabilities with its recent acquisition of database and machine learning vendor DataStax.
A better story
However, Shimmin said IBM needs to tell a better story about its operations and all the products it has that can be helpful to enterprises within the Watsonx Orchestrate system. He noted that IBM is also proficient in RPA technology, process mining and rules engines, yet the vendor has not effectively promoted its expertise in those areas..
And IBM might not be able to market the tools or integrate them smoothly in a way that’s manageable for its longstanding enterprise customers.
"It's one thing to have all the great technologies at hand to support this kind of complex efforts, but it's another if you can package them up in something consumable by both IT and the business," Shimmin continued. "That latter part is where I think IBM has a lot of work to do in showing their customers that they're a complete end-to-end platform easily consumable by the business and IT."
In addition to agentic capabilities and data tools, IBM launched IBM LinuxOne 5, its latest open source Linux platform. This version of the platform includes IBM's Telum II on-chip AI processor and the IBM Spyre Accelerator card for enabling generative and high-volume AI applications. The vendor said moving cloud-native, containerized workloads to IBM LinuxOne 5 running the same software products can save enterprises considerably.
Esther Shittu is an Informa TechTarget news writer and podcast host covering AI software and systems.
