IBM at its annual THINK 2025 conference today added a suite of domain-specific artificial intelligence (AI) agents that can be integrated using a framework, dubbed watsonx Orchestrate, that it is adding to its integration portfolio.
Via this orchestration framework, IBM will also provide observability capabilities for performance monitoring, guardrails, model optimization and governance across the entire agent lifecycle.
IBM is also making available a range of tools that enable organizations to build their agents in minutes, in addition to introducing webMethods Hybrid Integration, an integration framework based AI agents and application programming interfaces (APIs). IBM acquired webMethods last year.
IBM is also previewing an Agent Catalog based on watsonx Orchestrate that provides access to more agents and pre-built tools from both IBM and third-party partners, and updating its watsonx.data framework to integrate an open data lakehouse with existing data fabric capabilities such as governance and data lineage tracking. The company, for example, is adding a single-interface tool for orchestrating data across formats and pipelines, and watsonx.data intelligence to apply AI technologies for extracting insights from unstructured data. Those offerings come on the heels of IBM’s move earlier this year to acquire DataStax. IBM is committing to adding watsonx as an application programming interface (API) that can be incorporated into the Llama Stack AI platform developed by Meta.
Finally, IBM revealed that LinuxONE 5, a distribution of Linux optimized for its z17 mainframe configured with Telum II processors will be able to process 450 billion inference operations per day
In total, IBM at the launch of this agentic AI initiative is providing integrations with more than 80 enterprise applications from partners such as Adobe, Amazon Web Services (AWS), Microsoft, Oracle, Salesforce, ServiceNow and Workday. In addition, IBM now supports the Model Context Protocol (MCP) developed by Anthropic along with an open source BeeAI project based on the Agent Communication Protocol (ACP). IBM last month donated BeeAI along with two other related AI technologies to the LF AI & Data Foundation, an arm of The Linux Foundation.
Despite the current level of economic uncertainty, organizations continue to invest heavily in AI, says IBM chairman and CEO Arvind Krishna. In fact, an IBM CEO survey published today finds growth rate of AI investments to more than double over the next two years. “They are doubling down on AI,” he says.
However, the challenge is that only 25% of AI initiatives have achieved the return on investment (ROI) they expected, according to the IBM survey.
It’s not clear to what degree the rise of agentic AI might enable IBM to recover some of the ground it lost with the rise of generative AI technologies. While IBM has been a long-time AI pioneer, startups such as OpenAI with help from Microsoft have since become AI industry leaders. However, many of the initial use cases for generative AI did not easily incorporate enterprise data. IBM views that gap as a major opportunity, Krishna says.
The issue now will be convincing enterprise IT organizations that today have no shortage of integration options to choose to standardize on IBM tools, frameworks and platforms.
