IBM has announced a range of new hybrid AI tools and services at its THINK conference, aimed at addressing the challenges businesses face in deploying AI across complex technology environments.
The updates are part of IBM’s effort to support AI development using enterprise data, as the company responds to forecasts suggesting more than one billion applications will be created by 2028. This expected growth is increasing demand for tools that support integration and coordination across different systems.
A new IBM CEO study reports that company leaders expect AI investment to more than double in the next two years. However, only a quarter of AI initiatives have achieved the returns companies were aiming for, with fragmented systems and limited integration among the reported difficulties.
AI agents in watsonx Orchestrate
IBM is expanding its watsonx Orchestrate platform, which allows users to build, deploy and manage AI agents. The platform includes tools for both technical and non-technical users and offers a catalogue of more than 150 agents. These include tools for tasks such as identifying sales prospects in Salesforce or answering HR queries through Slack.
Agents can connect with more than 80 business applications, including Adobe, AWS, Microsoft, Oracle, SAP, ServiceNow and Workday. IBM is also adding features for performance monitoring, coordination between tools, and oversight of how agents are used.
Also read: Cognizant CEO says future tech will be driven by machines writing software
Hybrid Integration for application connectivity
To address integration between systems, IBM is introducing webMethods Hybrid Integration. The software allows users to manage application programming interfaces (APIs), file transfers, and other connections between software and cloud environments.
An independent study by Forrester Consulting found that companies using webMethods reported a 176% return on investment over three years. The study also found reductions in system downtime, shorter project timelines, and decreased training costs.
These capabilities will be used alongside IBM’s other automation tools and integrations with providers such as HashiCorp and Red Hat. The aim is to improve how systems are provisioned, maintained and scaled across different environments.
Unstructured data and watsonx.data
IBM is also expanding its watsonx.data product to support the analysis of unstructured data from documents, spreadsheets and presentations. The updates combine a data storage structure with features to manage, access and use data stored across various systems.
New tools under watsonx.data include a user interface for managing data pipelines and an AI system for extracting insights from documents. IBM’s internal testing suggests this can produce more accurate AI responses compared to traditional methods.
Also read: AI Mode in Google Labs now available without waitlist: Here’s what it can do
IBM recently announced plans to acquire DataStax to add search capabilities that support unstructured data analysis. IBM’s watsonx tools are also now available as part of Meta’s Llama Stack, a software suite for building and running large language models.
IBM has added content-aware storage to its Fusion service to support data processing. This allows extracted data to be used in AI systems more quickly.
Infrastructure for AI
IBM has launched LinuxONE 5, a Linux-based server designed to support data processing and AI workloads. The system includes IBM’s Telum II chip and a new accelerator card, which is due later this year.
LinuxONE 5 includes security features such as container isolation and encryption designed for future threats. IBM reports that running software on LinuxONE 5 can lower total costs by up to 44% over five years when compared to equivalent x86 systems.
IBM has expanded its collaborations with AMD, Intel, CoreWeave and NVIDIA to offer additional processing and storage options for companies using data-heavy applications.
Also read: A Chinese AI can look at pics of patients’ eyes and detect kidney disease in diabetics