Enterprise AI is booming, messy, and—more often than many leaders admit—dangerously inaccurate. OpenText thinks it knows why: organizations have unleashed AI on oceans of unstructured, unlabeled, poorly governed data, then act surprised when the models hallucinate, misinterpret, or leak sensitive information.

This week at OpenText World 2025, the company revealed its counterstrategy: the OpenText AI Data Platform (AIDP), an open, governed data layer engineered to give enterprise AI the one thing it consistently struggles with—context.

Where other vendors chase bigger models or flashier agents, OpenText is doubling down on its heritage: decades of document management, metadata discipline, and enterprise-grade information governance. In an era where half of AI-using organizations report at least one serious accuracy or risk failure (McKinsey’s numbers, not OpenText’s), the pitch hits close to home.

OpenText’s message is blunt: if the data is wrong, the AI will be wrong—no matter how impressive the model is.

Why Contextual AI Is the Hill OpenText Wants to Die On

OpenText has spent more than 30 years holding, securing, and classifying some of the world’s largest enterprise datasets. That experience underpins its thesis: AI agents only become useful when they understand where they are, what they’re allowed to see, and why a task matters.

Documents. Tickets. Commerce records. Security logs. Machine outputs. Human inputs.
All tagged, secured, governed, versioned, and compliant.

OpenText says enterprises must treat AI less like a chatbot experiment and more like a discipline rooted in data lineage, identity access control, retention policies, and contextual metadata. Otherwise, even the smartest models become highly efficient generators of confusion.

This foundation feeds directly into OpenText Aviator, the company’s enterprise AI engine, which can now orchestrate workflows through domain-aware agents.

The Three Pillars of Aviator’s Architecture

OpenText insists it’s not building another AI walled garden. Aviator’s architecture leans heavily into openness:

1. Multi-cloud
   Works across on-prem, cloud, hybrid, or multi-cloud deployments.

2. Multi-model
   Compatible with any LLM or SLM—including “bring your own model.”

3. Multi-application
   Built for deep integration with ERP, CRM, ITSM, security suites, and more.

In reality, this means OpenText wants its AI agents to plug into the daily arteries of enterprise work—from SAP order flows to Salesforce deals to Oracle records to Microsoft infrastructure.

“Everyone is chasing the mega-agent. But enterprises need armies of domain-specific agents,” said Savinay Berry, CPO & CTO at OpenText. “Accuracy through trusted data isn’t an IT feature—it’s a C-level mandate.”

The Databricks Partnership: A Strategic Data Power Play

A major announcement embedded in the platform launch is OpenText’s expanded partnership with Databricks. The companies will co-innovate on AIDP with deeper technical integrations, Delta Sharing, and a unified governance path.

OpenText already ran Threat Detection and Response on the Databricks Data Intelligence Platform. Now the partnership widens into joint engineering.

The intent is clear:
Combine Databricks’ analytics engine with OpenText’s governed data fabric to deliver trustworthy, enterprise-ready AI.

If successful, this pairing could become a serious contender against Microsoft’s Fabric, Google’s Vertex-BigQuery pipeline, and Snowflake’s AI-ready enterprise stack.

Inside the 18-Month Roadmap: OT 26.1 to OT 27.2

At OpenText World, the company revealed a surprisingly detailed roadmap for the next six releases:

OpenText AI Data Platform (AIDP)

A unified data and AI framework with governance orchestration. Think of it as a control tower for every agent decision.

OpenText Aviator Studio

A no-code environment for building and governing enterprise AI agents—without requiring data scientists to hand-craft pipelines.

OpenText Knowledge Discovery

A metadata-first ingestion engine that transforms structured and unstructured data into AI-ready context.

OpenText Data Compliance

A suite spanning privacy, tokenization, encryption, PII controls, redaction, AI readiness checks, and threat detection.

OpenText Aviator AI Services

A professional services track to help enterprises move from AI experiments to production-grade agent deployments.

This aggressive roadmap signals OpenText’s belief that the battle for enterprise AI will be fought not in the model layer, but in the data and governance layer.

What Enterprises Can Use Today

OpenText emphasized that Aviator is already live for real-world use cases like:

• fraud detection

• claims management

• predictive maintenance

• customer service automation

• IT operations workflows

The company also announced that the Aviator entry-tier package will be included at no extra cost with upgrades to OT 26.1 for Content Management, Service Management, and Communications Management.

Better yet for risk-averse industries, Aviator will become fully available on-premises starting with OT 26.1 across multiple modules, including DevOps and Application Security.

For global enterprises navigating sovereignty laws, this on-prem push is a quiet but important differentiator.

What This Means for the Enterprise AI Landscape

OpenText is staking out a clear and contrarian position:
AI models do not matter unless the data behind them is governed, contextual, and trustworthy.

This philosophy diverges sharply from model-first players—hugging the foundational layers of enterprise information instead of competing in the model arms race. With model commoditization accelerating, that may prove to be a winning angle.

AIDP also signals a broader industry shift toward:

• governed AI pipelines

• enterprise-grade agent orchestration

• model-agnostic architectures

• contextual knowledge layers

• compliance-integrated design

In short, OpenText is rewriting AI around the data source, not the model endpoint.

If other vendors follow, the next generation of enterprise AI may finally behave less like an unpredictable intern and more like a dependable colleague.

Get in touch with our MarTech Experts.