IBM Content Cortex
AI-powered enterprise content management
IBM Content Cortex is IBM’s next-generation content management platform. It combines decades of experience from FileNet, Content Manager and Content Manager OnDemand in a single, AI-ready system. On the platform, AI agents can work directly in the enterprise document repository and carry out tasks that typically require manual effort today, while complying with existing access control and retention rules.
The evolution of Content Cortex: from three product families to one platform
For a long time, IBM’s content management portfolio was built on three parallel product families. FileNet managed active documents and processes, Content Manager served enterprise document repositories, and Content Manager OnDemand (CMoD) handled high-volume archived output such as statements and invoices. As a result, many organizations still run two or three systems side by side, each with separate operations and separate licenses.
IBM announced the unification of the three product families at its Think 2026 conference; the resulting IBM Content Cortex has been available since June 2026. The platform is built on FileNet’s proven, enterprise-scale foundations, but its architecture was designed from the outset for AI agents. The core of IBM’s strategy is that the bottleneck in adopting AI is usually not the language model, but a fragmented, poorly indexed document estate. Content Cortex therefore aims to turn the document repository from a passive archive into a trusted content foundation on which automation can be built safely.
The transition can be carried out gradually: existing FileNet licenses can be carried over, and Content Manager and CMoD content can be moved through a scheduled migration. Planning the transition is also timely because general support for IBM FileNet Content Manager 5.5.x ended on September 30, 2026; from then on, it is available only as extended support at an additional cost. According to IBM’s current schedule, versions 5.6 and 5.7 are supported until September 2028.
Technology and new capabilities
Content Cortex runs as a containerized application on the Red Hat OpenShift platform and can be deployed on premises, in a private cloud or in a public cloud with the same functionality. Availability and scaling are handled by standard OpenShift tools, and the entire configuration can be stored in Git and deployed through a CI/CD pipeline. The platform is available in Essentials and Premium editions.
The most important new capabilities are:
- AI agents in the document repository: Navigator Core Agent searches based on natural-language instructions, fills in metadata from the document’s content (quoting the source text next to each value), and reclassifies and files documents. Before making any change, it presents the planned steps and carries them out only after approval. The Premium edition also includes redaction, legal hold, and classification and extraction agents.
- Question answering and comparison: Content Assistant creates summaries and answers questions based on the documents, referencing the relevant section. When comparing two versions or a resubmission of missing documents, it lists the significant differences, while the two documents can also be viewed side by side.
- Unified search: Metadata, full-text and semantic (vector) search in a single query, with search templates that can be saved and shared.
- Open integration: In addition to CMIS and GraphQL interfaces, the platform provides MCP (Model Context Protocol) servers, so external AI platforms such as IBM watsonx Orchestrate or Microsoft Copilot can access content under the same access control and audit rules.
- Choice of language models: The AI features work with IBM watsonx services or with open models running on the organization’s own GPUs. On our own GPU infrastructure, we run the entire AI layer locally, so documents never leave the organization.
Business benefits and use cases
The business value of Content Cortex shows mainly in processes where documents are still handled manually today. For incoming applications, invoices or requests from authorities, AI performs classification and data extraction, so case handlers only need to review the exceptions, and the changes in a resubmission can be reviewed in a few minutes. New document types, fields and templates can be introduced in days through configuration instead of a development project.
Meeting compliance requirements is built into the platform. The system logs every event, including the steps taken by AI agents, applies certifiable retention and disposition rules, and monitors the expiry of electronic signatures for timely re-signing. GDPR erasure covers metadata, content and AI-generated indexes alike, and confidential data can be redacted based on role. All of this runs on a single standard platform instead of the previous two or three systems, which reduces operating costs and risk.