IBM said on February 17, 2026 that it is bringing AI directly into the software systems enterprises already use to run data, analytics, messaging, engineering, supply chain, and integration workloads. Rather than asking customers to replatform around a new AI stack, IBM is embedding agentic AI into existing products with the stated goal of simplifying operations, accelerating insight, and improving governance inside regulated environments.

The announcement covers a broad first-quarter rollout. IBM says new AI capabilities are planned for Db2 on March 5, Sterling Order Management System on March 6, Cognos Analytics on March 12, MQ on March 24, Cloud Pak for Integration on March 25, Engineering Lifecycle Management on March 26, and both Sterling B2B Integration SaaS and App Connect Enterprise on March 31.

Where IBM is adding agents

In Db2, IBM is introducing a large set of operational agents, including AskDb2, AskDb2 Research Agent, Telemetry Agent, Text to SQL Agent, Change Guard, Spill Guard, Queueing, and several others aimed at performance tuning and troubleshooting. In Cognos Analytics, IBM is adding reporting agents that recommend reports, summarize them, and automate secure sharing. MQ and Cloud Pak for Integration are both getting Supervisor and Knowledge agents along with specialized agents for topology, logs, versions, and operational diagnostics.

IBM is also extending agentic workflows into industry and engineering software. Sterling OMS is getting an Agentic Toolkit Add-On with order, inventory, cancellation, coupon, and contract risk agents. Engineering Lifecycle Management is adding a Work Item compose agent inside Engineering Workflow Management. Sterling B2B Integration SaaS is adding InFlight orchestration with document, event, and anomaly-oriented agents. App Connect Enterprise is introducing agents for deployment, topology analysis, flow inspection, knowledge retrieval, and resource management.

Why the portfolio move matters

The important part of this release is its scope. IBM is not announcing one assistant attached to one application. It is spreading domain-specific agents across a large portfolio of operational systems where enterprises already store data, manage business rules, route messages, and run integrations. That gives IBM a path to make AI useful through existing workflows rather than through isolated demos.

The company is also leaning heavily on governance. The announcement repeatedly frames the rollout around security, compliance, and trust in complex environments. That positioning matters for sectors where AI adoption is constrained not by lack of interest but by the difficulty of connecting new automation to established systems without losing control. If IBM can deliver these agents on the timetable it outlined for March 2026, the company will have moved agentic AI from slideware into some of the enterprise software categories where operational reliability matters most.

#enterprise-software

IBM embeds agentic AI across Db2, Cognos, MQ, and integration software in Q1 2026