IBM Unveils Tools for AI Agents and Enterprise Data

IBM ramps up its efforts to push AI deeper into the enterprise by expanding its Watsonx platform and unveiling new integration and data tools that aim to address one of the technology’s biggest obstacles: delivering measurable value.

The announcement, made during the company’s Think event, outlines IBM’s strategy to help organisations design and deploy AI agents that integrate with complex digital environments and fragmented data sources.

Building AI agents for enterprise tasks

IBM says a key issue hindering enterprise AI is the failure to integrate systems across hybrid and multi-cloud setups.

The company’s internal research shows that although interest in AI continues to grow, only 25% of AI initiatives have met expectations in terms of return on investment.

IBM points to technology sprawl and a lack of orchestration as reasons for this shortfall.

To tackle this, IBM expands Watsonx Orchestrate, its AI agent-building platform.

The system now includes tools for creating custom AI agents in less than five minutes.

These agents, designed to perform tasks across sales, procurement and HR, can be built without coding or within professional development environments.

This flexibility is meant to support users of varying technical skill levels.

Watsonx Orchestrate now also hosts an Agent Catalog, a library with more than 150 pre-built tools and agents from IBM and its partners.

These include names like Salesforce, Oracle, ServiceNow, Box and Symplistic.ai.

The platform connects with over 80 enterprise applications, ranging from CRM and cloud providers to business operations platforms.

These AI agents are described by IBM as modular tools that can automate tasks such as document handling, data lookup and even decision-support operations.

They work by processing enterprise data and triggering actions through APIs or predefined workflows.

“The era of AI experimentation is over,” says Arvind Krishna, IBM’s Chairman and Chief Executive.

“Today’s competitive advantage comes from purpose-built AI integration that drives measurable business outcomes.”

Tackling integration complexity with hybrid AI technology

Another key component of IBM’s announcement is webMethods Hybrid Integration.

This solution is intended to replace traditional workflow automation by enabling agent-driven integration across systems, including APIs, cloud environments and partner ecosystems.

IBM says that as organisations face increasing pressure to manage sprawling tech infrastructure, hybrid integration allows AI agents to automate and coordinate processes that span across digital platforms.

This is especially relevant in hybrid cloud settings, where business data and services are often distributed.

According to a Forrester Consulting study commissioned by IBM, customers using webMethods tools experienced a 176% return on investment over three years.

The report also shows a 40% drop in system downtime, as well as major time savings on both complex and routine tasks.

These results suggest that integrating AI agents at the infrastructure level can create practical benefits that justify the investment.

IBM also expands its automation capabilities through deeper partnerships, including enhanced support for tools like HashiCorp Terraform and Vault.

These integrations are designed to improve deployment workflows and secrets management across hybrid setups.

Managing unstructured data for smarter AI outcomes

Beyond integration, IBM turns its attention to one of AI’s persistent challenges: unstructured data.

From contracts and spreadsheets to slide decks and chat logs, unstructured formats make up the majority of enterprise information.

Watsonx.data, IBM’s AI data platform, is being updated to tackle this issue directly.

The platform combines an open data lakehouse architecture with what IBM calls data fabric capabilities.

This includes data lineage tracking and governance, which help organisations bring consistency to datasets scattered across cloud and on-premise systems.

According to IBM, using watsonx.data to power AI agents leads to results that are 40% more accurate than traditional retrieval-augmented generation models.

Two new tools expand this capability further.

Watsonx.data integration allows businesses to manage data pipelines from a single interface, while watsonx.data intelligence applies AI to extract insights from unstructured sources.

Both tools are being offered as standalone products and will also be partially available within the main watsonx.data platform.

In a separate move, IBM plans to acquire DataStax, a company known for helping Gen AI tools manage and query unstructured data.

IBM says this deal will provide clients with access to vector search capabilities — a function that enables AI agents to find relevant data even if it’s not structured or labelled in a traditional format.

IBM’s integration work also touches Meta’s Llama Stack.

The company confirms watsonx has now been added as an API provider, helping businesses deploy large language models in enterprise-grade environments.

Looking ahead, IBM also launches LinuxONE 5, which it calls its most secure Linux system.

This platform is designed for high-volume AI and data workloads and is powered by the Telum II processor and the IBM Spyre Accelerator.

It is expected to be available by the fourth quarter of 2025 and will support up to 450 billion AI inference operations per day.

“IBM is equipping enterprises with hybrid technologies that cut through complexity and accelerate production-ready AI implementations,” Arvind says.

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today. 

AI Magazine is a BizClik brand