At its annual THINK event, IBM launched a suite of cutting-edge hybrid AI technologies designed to help enterprises scale Generative AI (Gen AI) across fragmented IT environments. With more than one billion new applications projected by 2028, IBM aims to provide the integration and orchestration tools needed to build, deploy, and scale AI agents effectively.
According to a new IBM CEO study, business leaders expect AI investments to more than double within two years. Yet only 25% of current AI initiatives have achieved their expected return on investment, largely due to disconnected technologies. In response, IBM is leveraging its hybrid cloud infrastructure, watsonx AI platform, and IBM Consulting services to address these challenges head-on.
AI Agents That Work Across 80+ Enterprise Applications
A major highlight of IBM’s announcement is the enhancement of watsonx Orchestrate, which now enables businesses to create AI agents in under five minutes using no-code to pro-code tools. These agents can operate seamlessly across more than 80 enterprise applications — including those from Microsoft, Adobe, Salesforce, Oracle, AWS, and ServiceNow.
IBM is also launching a new Agent Catalog within watsonx Orchestrate, featuring over 150 pre-built agents and tools co-developed with partners like MasterCard, Box, Oracle, Symplistic.ai, and 11x. For example, a sales agent integrated with Salesforce Agentforce and a conversational HR agent compatible with Slack are among the first offerings.
WebMethods Hybrid Integration Delivers 176% ROI
To support seamless AI integration across hybrid cloud environments, IBM introduced webMethods Hybrid Integration. This next-generation automation solution helps businesses manage app, API, file, and partner integrations across cloud and on-premises systems. An independent Forrester TEI study projected a 176% return on investment over three years, including a 67% time savings on simple projects and a 40% reduction in downtime.
Unlocking the Value of Unstructured Data
IBM is also addressing one of AI’s biggest bottlenecks: unstructured data. The new watsonx.data update combines an open data lakehouse with powerful governance and data fabric capabilities, allowing AI applications to tap into contracts, spreadsheets, and presentations. Early testing suggests this approach can boost AI accuracy by up to 40% compared to traditional retrieval-augmented generation (RAG) methods.
In addition, watsonx.data integration and watsonx.data intelligence, both offered as standalone or integrated solutions, help businesses manage and extract insights from diverse data sources. IBM’s pending acquisition of DataStax will further enhance vector search capabilities, while watsonx’s integration into Meta’s Llama Stack expands generative AI flexibility and openness.
Infrastructure Designed for AI at Scale
To support the massive computational demands of Gen AI, IBM introduced the IBM LinuxONE 5 — its most powerful and secure Linux platform yet. It can handle up to 450 billion AI inference operations per day and features next-gen AI accelerators, including the IBM Telum II chip and the forthcoming Spyre Accelerator. IBM also promises up to 44% total cost savings over five years compared to comparable x86 systems.
Rounding out the announcement, IBM revealed expanded collaborations with AMD, Intel, CoreWeave, and NVIDIA to deliver enhanced compute and storage solutions for AI-intensive workloads.
With these innovations, IBM is positioning itself as a leader in enabling scalable, secure, and enterprise-ready generative AI, ushering in what it calls the end of AI experimentation and the beginning of measurable, production-grade outcomes.