IBM got here out swinging at its annual THINK 2025 occasion this week, asserting a wave of contemporary AI capabilities, deeper partnerships, and sturdy new infrastructure that alerts the corporate goes all-in on enterprise AI.
From five-minute AI brokers to a hybrid cloud revolution, IBM says it’s time to cease dabbling in synthetic intelligence and begin delivering outcomes.
“The period of AI experimentation is over,” stated IBM Chairman and CEO Arvind Krishna in the course of the occasion. “At the moment’s aggressive benefit comes from purpose-built AI integration that drives measurable enterprise outcomes.”
AI brokers that work, not simply speak
IBM is positioning its new suite of agent capabilities in Watson Orchestrate because the spine of the subsequent technology of AI within the office. Not like conventional chatbots, these new brokers aren’t simply conversational – they’re constructed to behave.
Want an agent to seek out gross sales prospects in Salesforce, deal with HR questions in Slack, or run procurement workflows in Oracle? IBM says you may construct that agent in underneath 5 minutes utilizing no-code or pro-code instruments. The Watson Orchestrate platform connects with greater than 80 enterprise purposes, together with Microsoft, ServiceNow, and Workday.
To simplify deployment, IBM can be launching an Agent Catalog, that includes greater than 150 pre-built brokers and instruments from companions like Field, Mastercard, and Symplistic.ai.
Oracle + IBM: A brand new AI energy duo
In a significant partnership growth, IBM is teaming up with Oracle to carry Watsonx Orchestrate to Oracle Cloud Infrastructure (OCI) beginning in July. This integration goals to streamline how firms use AI throughout Oracle and non-Oracle apps.
“AI delivers essentially the most impactful worth when it really works seamlessly throughout a complete enterprise,” stated Greg Pavlik, government vice chairman of AI and Information Administration Companies at Oracle Cloud Infrastructure. “Our expanded partnership will present prospects new methods to assist rework their companies with AI.”
Below the deal, IBM’s Granite AI fashions can even be obtainable in Oracle’s AI instruments, and watsonx.ai is now licensed to run on OCI, permitting for AI deployments nearer to enterprise information and purposes.
Hybrid integration for the AI age
Integration stays one of the vital difficult hurdles for enterprise AI. To handle this, IBM is introducing webMethods Hybrid Integration, a unified platform that connects APIs, apps, occasions, information, and even mainframe information throughout hybrid cloud environments.
SEE: New IBM z17 Mainframe Will ‘Redefine AI at Scale’
IBM’s hybrid integration method leverages AI to automate the method, from producing integrations with pure language to managing APIs and B2B information flows.
In keeping with a Forrester Consulting research, firms utilizing IBM’s webMethods instruments achieved a 176% return on funding (ROI) over three years, reduce downtime by 40%, and noticed a 67% time financial savings on easy initiatives.
Turning unstructured information into AI gasoline
IBM’s new watsonx.information updates purpose at a persistent enterprise downside: unstructured information. Suppose contracts, spreadsheets, shows — all of the stuff hiding in folders that conventional techniques ignore.
The up to date platform introduces information lakehouse structure, AI-powered insights, and simpler connections to AI brokers. IBM claims the enhancements can increase AI accuracy by as much as 40% in comparison with standard retrieval-augmented technology (RAG) strategies.
To additional this mission, IBM is buying DataStax, recognized for its unstructured information and vector search capabilities, and integrating with Meta’s Llama Stack for open AI deployment.
A brand new engine for enterprise AI: LinuxONE 5
Behind all these AI instruments lies severe computing energy. IBM launched LinuxONE 5, its next-generation Linux server constructed for enormous AI workloads, able to operating as much as 450 billion inference operations per day.
This powerhouse combines safety, power effectivity, and scalability. IBM says switching from x86 to LinuxONE 5 might reduce complete price of possession by as much as 44% over 5 years. It additionally contains built-in AI accelerators and quantum-safe encryption, which goals to “deal with quantum-enabled cybersecurity assaults.”
Agentic AI throughout the enterprise
From consulting providers to integration platforms, IBM is wrapping all of its strikes into an “agentic AI” technique. IBM Consulting helps prospects undertake this mannequin, together with migrating legacy workloads to Crimson Hat OpenShift Virtualization on OCI.
“Brokers have gotten an essential strategic benefit for organizations to optimize operations and ship higher buyer and worker experiences,” stated Ritu Jyoti, group VP at IDC, as quoted within the IBM/Oracle press launch. “The Agentic AI method IBM is taking with Oracle is a number one instance of how the flexibility to orchestrate agentic workflows throughout techniques can drive vital benefits.”