‘One of the biggest reactions we’ve gotten on Watsonx.orchestrate and agents is from services partners. They want to build agents. They want a way to have them land on a platform like Orchestrate, where they can leverage multiple models, multiple data repositories,’ says IBM Senior Vice President of Software and Chief Commercial Officer Rob Thomas.
New domain-specific artificial intelligence agents. Advancements in AI orchestration tools. And details on the latest version of IBM’s LinuxOne.
These are some of the biggest announcements coming out of the Armonk, N.Y. cloud, mainframe and AI vendor’s annual Think conference. Think 2025 runs through Thursday in Boston.
During a virtual press conference, IBM Senior Vice President of Software and Chief Commercial Officer Rob Thomas told CRN that solution providers are “a critical part of how clients acquire and consume technology” and represent “one of the biggest investments we’ve made” over the last four years.
He put services at north of 80 percent of the $5 trillion a year in projected IT spending and credited solution providers with Watsonx.orchestrate adoption.
“One of the biggest reactions we’ve gotten on Watsonx.orchestrate and agents is from services partners,” Thomas said. “They want to build agents. They want a way to have them land on a platform like Orchestrate, where they can leverage multiple models, multiple data repositories. And so we’ve got a lot of momentum with systems integrators building agents using Orchestrate.”
[RELATED: What IBM’s Earnings Say About Tariffs, Consulting, AI]
IBM Think 2025
In response to another question during the conference, IBM Chairman and CEO Arvind Krishna said that AI demand has not fallen amid economic uncertainty as countries negotiate global tariffs.
“Everybody is doubling down on AI investments, but they’re now looking for that return on AI,” he said. “The only change over the last 12 months is that people are stopping experimentation and focusing very much on where is the value to the business right now?”
Perhaps of importance to solution providers, Krishna said that consultancies and services businesses can leverage new AI software to create platforms from previously conducted knowledge work, using EY’s tax question-and-answer product built on Watsonx.ai as an example.
“Now their consultants can be more productive,” he said. “It actually allows them to scale also in going more downmarket because if the cost of providing this consulting is now cheaper, you can scale to many smaller clients. … Once the platform is built out, it offers them cost savings.”
Heather Domin, vice president and head of office for responsible AI and governance at India-based IBM partner HCLTech, told CRN in an interview that IBM investing in not just AI applications but the underlying technology making AI usable for enterprises is important for her practice.
She cited the Watsonx.governance platform and Guardium real-time activity monitoring product.
“We are seeing a lot of interest with our clients around responsible AI and, of course, IBM brings just a tremendous amount of expertise in that space,” said Domin, who previously served as IBM’s global leader for responsible AI initiatives. “One of the things that I love about IBM is that there really is a strong awareness of where the technology is heading in general, and then what are the types of guardrails that we need to have in place. And I’d love to see that in all partners and vendors that we work with.”
The news helps illustrate the ways IBM and its partners are going after the $4.4 trillion annual economic value that GenAI is expected to generate across industries.
AI is also the way to scale across an estimated 1 billion-plus applications that will be in existence by 2028, according to IBM. That many apps means countless fragmented environments.
Data that IBM shared during Think shows that CEOs expect the AI investment growth rate to more than double over the next two years. But so far, only 25 percent of AI initiatives have achieved the expected ROI.
Read on for more of IBM’s biggest news from Think 2025.
Watsonx Domain Agents
IBM Watsonx human resources agents are available and aimed at automating employee support workflows, from time-off management and profile updates to leave and benefits, according to the vendor. These agents integrate with Workday Human Capital Management and other HR systems.
Now available in public preview with general availability planned for June are Watsonx Procurement agents and Watsonx Sales agents. Procurement agents are meant to address procure-to-pay, supplier assessments, vendor management processes and other similar workflows. These agents also integrate with Coupa, Sirion, Dun & Bradstreet and other procurement tools.
Watsonx Sales agents support new prospect identification, qualified leads outreach, research and enablement, and other sales processes. The agents integrate with Salesforce, Seismic and other sales technologies.
Throughout the year, IBM plans to release domain agents in customer care, finance and other areas. IBM also has plans for general utility agents for web research, calculations and similar tasks. These agents can embed within users’ existing systems.
Watsonx Orchestrate Updates
In June, IBM will release a series of new capabilities for Watsonx Orchestrate users, including the ability to build agents in under five minutes, agent observability for performance monitoring and an agent catalog.
Watsonx Orchestrate Agent Builder is a no-code automation studio that promises the ability to build and customize agents in less than five minutes, according to IBM. The agents are grounded in business knowledge and can leverage prebuilt tools, making the agent builder available for highly skilled coders and users with no code experience.
Agent observability will allow users to monitor agent performance, optimize models and apply guardrails and governance across the agent life cycle.
The Watsonx Orchestrate Agent Catalog introduced during Think aims to simplify access to IBM agents and ones offered by third parties, according to the vendor. The catalog is how users can access hundreds of prebuilt tools and agents.
A new Watsonx Orchestrate program allows any IBM partner to build directly with the product to help expand the portfolio of third-party applications, according to the vendor. Expanding IBM agent connectivity with support for Anthropic’s Model Context Protocol also fuels IBM agents’ integration capabilities.
During Think, Orchestrate also gained multiagent orchestration capabilities to empower agent-to-agent work.
Agents built in-house by partners or with open source can share information and address multistep processes. Users can embed multiagent capabilities into agentic systems. User requests are analyzed and routined across agents, assistants and skills in real time.
Watsonx Orchestrate should enable agent collaboration leveraging existing automations, APIs, data sources and core apps.
As part of IBM’s Think announcements, the vendor revealed a new agent development kit for building specialized agents from scratch and deploying from laptops into Watsonx Orchestrate.
Watsonx.data Updates
June is also IBM’s planned availability date for a new version of Watsonx.data that unifies an open data lakehouse with data lineage tracking, governance and other data fabric capabilities.
Users will have the ability to leverage Watsonx.data to bring together, govern and activate data across silos, formats and clouds. Users can connect AI apps and agents with unstructured data and receive more accurate outputs. IBM put Watsonx.data’s AI at 40 percent more accurate than the conventional retrieval-augmented generation framework.
June will also mark availability for two stand-alone Watsonx.data products, with some capabilities available through the larger data lakehouse—Watsonx.data integration and Watsonx.data intelligence.
Integration is a single-interface tool for orchestrating data across pipelines and formats. Intelligence leverages AI for finding insight in unstructured data.
WebMethods Hybrid Integration
IBM will make WebMethods Hybrid Integration available in June for users looking for a way to manage integration sprawl in the AI era.
IBM positions WebMethods Hybrid Integration as a way to bridge the hodgepodge of APIs, applications, systems, partners, events, gateways, file transfers and more integrations in hybrid cloud, multi-cloud and on-premises environments.
The adoption of multiple WebMethods integration capabilities brought some customers redacted downtime, time savings on projects and improved visibility and security posture, according to IBM.
IBM LinuxOne 5
IBM said its next version of LinuxOne promises the most secure Linux platform for AI, with the ability to process 450 billion-plus inference operations a day.
LinuxOne 5 can leverage IBM’s Telum II on-chip AI processor and IBM Spyre Accelerator card—–slated for PCIe card release in 2025’s fourth quarter—for transactional workloads and other high-volume AI apps.
The new Linux platform brings in confidential containers and other advanced security offerings. It interacts with IBM quantum-safe encryption technology and promises reductions in costs and power consumption.
IBM said that compared with x86 platforms, LinuxOne 5 can save up to 44 percent in TCO over five years.