OpenText World 2026, held October 19–22 at the Hyatt Regency Orlando, put context at the center of the enterprise AI discussion. It was Ayman Antoun’s first OpenText World as CEO, and across keynotes, labs, and customer sessions, the company framed secure information management as the foundation enterprises need to turn AI ambitions into working systems.
The agenda’s seven tracks ran from AI-ready content and agentic AI governance through IT operations, customer experience, and software delivery to running OpenText at scale.
Key Announcements
Enterprise Content Moves Toward Agentic Execution
OpenText’s content portfolio is moving from AI assistance toward agents that act on enterprise information. Content Aviator gained its first agentic capabilities in the June 26.2 release of OpenText Content Management, and a hands-on lab introduced attendees to Aviator Studio for building agents on trusted enterprise content. Analyst firm Deep Analysis counted more than 150 Aviator agents available at the time of the show.
Useful enterprise agents need governed access to information with business context. OpenText named fragmented systems, inconsistent metadata, and disconnected processes as the constraints on what those agents can accomplish, which puts information architecture squarely into AI execution.
AI-Ready Information Becomes an Operational Requirement
OpenText devoted a session to Knowledge Discovery, which shipped its 26.3 release in August. The product uses intelligent enrichment, automated classification, and knowledge-graph generation to give agentic AI structured context for retrieval and action.
OpenText describes its broader AI strategy as connecting structured and unstructured data across ERP, CRM, content management, and cloud applications, then applying Aviator agents to the workflows built on that data.
Governance Gets Pulled Into the Agent Conversation
Security and control were built directly into the agenda through a track focused on governing and controlling agentic AI. Elsewhere, sessions connected information readiness with governance and emphasized that enterprise agents require trusted information plus appropriate controls before they can take meaningful action.
As organizations give AI systems operational responsibility, controlling what information agents can retrieve, interpret, and act upon becomes part of the security architecture.
Autonomous Operations Spread Beyond Content
The agent theme reached well beyond content management. Conference tracks addressed proactive IT operations, autonomous operations emerging from enterprise data, and autonomous software delivery that allows teams to accelerate releases while maintaining trust in the process.
OpenText expects AI to work inside operational workflows across large organizations. Deep Analysis reported the company closing notable deals in IT asset, service, and operations management, areas outside its traditional reputation.
Strategic Insights
Context Is Becoming AI Infrastructure
OpenText describes enterprise AI as dependent on a “secure and trusted context layer” to scale. The phrase puts the quality, governance, and relationships within enterprise information on the same footing as the AI models themselves.
Alongside model selection, organizations have to decide which content agents can access, how information is classified, how access rights follow users and agents, and whether source information carries enough business meaning to support reliable actions.
Agents Raise the Stakes for Security
The conference paired agentic execution with governance rather deliberately. One of the seven primary tracks addressed governing and controlling agentic AI, while content sessions stressed security, organizational controls, and trusted information as prerequisites for action-oriented systems.
OpenText’s 2026 Cybersecurity Threat Report, drawn from telemetry across tens of millions of endpoints, found that credential theft had become the fastest path to access and that spear phishing now accounts for the majority of phishing activity.
AI Value Gets Pushed Closer to Work
An AI lab walked attendees through building agents for OpenText Content Management, and an IT director from a high-tech company described how OpenText Managed Services automated an invoicing process that supports $4 billion in annual revenue. AI office hours with forward-deployed engineers and hands-on coding meetups kept the focus on implementation.
Manufacturing sessions addressed AI in supply chains and B2B data visibility, while public-sector sessions examined AI-ready content, automated classification, and knowledge graphs for more reliable agent responses.
OpenText Leaves Orchestration to Others
Shannon Bell, EVP and chief product and technology officer, said in Orlando that OpenText will build composable AI on its own products and plug into other vendors’ orchestration engines through industry-standard protocols. Last year, OpenText pitched its AI Data Platform as the orchestration layer for agents from multiple vendors.
Alongside CEO Ayman Antoun and Bell, featured speakers represented Microsoft, Amazon Web Services, Google Cloud, SAP, IBM, Deloitte, Gartner, and IDC.
Enterprise information sits across many applications and clouds, so OpenText’s agents have to work across those boundaries and answer to whichever orchestration platform a customer chooses.
The Undercurrent
AI readiness at OpenText World 2026 came down to information management. OpenText kept returning to trusted context, classification, and controlled access as the machinery behind credible enterprise AI.
The agenda promoted agents that execute content workflows and speed up software delivery, and it gave equal weight to governance and trusted data. Bell said customers asked for tactical agents that take time-consuming tasks off employees’ plates, which shows how much authority they are ready to grant an AI system today.
The Information Layer Sets the Agent’s Limits
AI programs that expect agents to perform business actions safely depend on unglamorous work in identity, permissions, metadata, and classification.
Organizations evaluating agentic systems should map the information each agent needs, identify the systems containing it, define permissions before deployment, and decide how actions are monitored, all before an AI pilot goes live.
Once an agent moves from answering a question to initiating an action, security teams, information owners, and business process owners all share responsibility for what it does.
What’s Next
Tech leaders planning for 2027 should review their information architecture alongside their AI roadmap. OpenText World 2026 made the connection explicit through its emphasis on trusted context, governed agents, and knowledge discovery.
Teams should begin with agent use cases where information access, permissions, and outcomes can be defined clearly, then test whether the underlying data and controls can support them. OpenText drew its own first scope the same way, building tactical agents inside its products and handing enterprise-wide orchestration to the platforms customers already use.