From “blow it up” to existing stack integration
AI agents deployment in the enterprise is the practice of connecting autonomous or semi-autonomous AI workflows directly into existing systems, data, and controls so they can act on business processes without requiring a full replacement of current technology stacks. Early AI vendors promoted standalone agent platforms that required fresh infrastructure, new data lakes, and sweeping process redesigns. That rip-and-replace pitch is losing steam. Enterprise leaders have learned that tearing out proven systems slows adoption, increases risk, and conflicts with established AI governance frameworks. Hyland CEO Jitesh Ghai calls the old playbook “blowing things up” and argues it is “improper” for organizations that already run complex content and process platforms. Instead, the emerging pattern is existing stack integration: meeting the organization where it is, adding context engines and meshes on top, and letting agents act through systems people already trust.
Context engines and meshes: agents that understand the enterprise
The new focus is less on stand-alone agents and more on how they gain context from enterprise AI infrastructure. Vendors now compete to be the context layer that feeds agents with content, metadata, and business rules. Hyland’s Enterprise Context Engine and Enterprise Agent Mesh show this shift clearly. Rather than forcing customers to move data or rebuild workflows, they sit across existing systems and content repositories, including unstructured documents that Ghai estimates make up 70% to 90% of enterprise data. By automating the “human ETL” between document and decision, agents can read, interpret, and act without people copying data between systems. This pattern treats content platforms, CRMs, and DXPs as durable assets, not obstacles, and turns them into shared context hubs that multiple agents can securely tap in real time.
Time-to-value and governance: why the existing stack wins
For IT and business leaders, existing stack integration is not just a technical choice; it is a time-to-value and governance decision. Reusing mature access controls, audit trails, and data policies means AI agents deployment can move from pilot to production without rebuilding compliance from the ground up. Liferay AI Hub is a clear example: it builds on Liferay DXP’s security framework so agents operate on behalf of authenticated users and can only access data those users are allowed to see. “The typical enterprise governance foundation includes access controls, data policies, and security infrastructure that have taken years to assemble,” notes Liferay’s Julia Molano. Similar thinking appears in Hexaware’s Agentverse, which adds policy-aware connectors, role-based access controls, and observability dashboards. In all cases, AI governance frameworks sit inside existing controls instead of standing beside them as a parallel stack.
How platforms are pivoting toward in-stack AI agents deployment
Leading platforms are quietly rewriting their roadmaps around in-place AI agents. Hyland extends its Content Innovation Cloud with a headless mode so agents can interact with its services directly, while its Enterprise Agent Mesh coordinates how agents talk to each other and to content systems. Hexaware’s Agentverse introduces Agentic Studios, a guided workflow—Define, Design, Approve, Test, Deploy, Operate—that plugs into Azure, AWS, and other infrastructures instead of competing with them. Liferay AI Hub takes a low-code approach, letting teams define and deploy agents grounded in existing DXP data and policies. In each case, the vendor’s value lies in orchestration, lifecycle management, and context—not in replacing core business platforms. This pivot reflects customer demand for predictable integrations that respect current investments and organizational change capacity.
Scaling AI with confidence means respecting what already works
Enterprises that want AI at scale are finding that the safest path is not a fresh, isolated AI stack but AI agents woven into what already works. Hexaware positions Agentverse as a “secure and high-performance” foundation, with lifecycle management and policy-aware connectors designed to keep agents aligned to business and compliance goals. Liferay AI Hub stresses that agents are “grounded in an organization’s own data and governed by its existing security policies,” reducing fragmentation and duplicate tooling. Hyland’s strategy of content-powered agents seeks to cut the 20% to 40% of knowledge workers’ time that Ghai says is lost to document-centric admin work. Across these examples, the pattern is clear: credible enterprise AI infrastructure keeps agents close to existing data policies and AI governance frameworks instead of asking CIOs to bet on a clean slate.
