From AI Features to Enterprise AI Agents
Enterprise AI agents are software entities that can interpret context, call systems and APIs, and take operational actions on behalf of users under defined governance and audit controls, turning AI from isolated features into workflow-ready digital workers embedded across business applications and platforms. The latest moves from OpenText and SAP show that these agents are no longer side experiments. OpenText is committing €105 million to expand operations and R&D around agentic AI, sovereign cloud infrastructure, and cybersecurity. SAP, in parallel, has announced a €100 million partner fund to drive production use of SAP Business AI and Joule-based agents. Together, these commitments signal a pivot: large application vendors are building agentic AI platforms and incentives for AI agent deployment, not only sprinkling generative AI across interfaces. For CIOs and enterprise architects, the question becomes how to plug these agent platforms into existing stacks without losing control of data or process integrity.
OpenText Bets on Sovereign Cloud Infrastructure Plus Agents
OpenText is investing €105 million in Cork and Galway over three years to expand what it calls foundational capabilities for trusted enterprise AI: agentic AI, sovereign cloud, and cybersecurity. Irish-based teams will design and operate multi-agent collaboration, system-boundary enforcement, and knowledge sharing across sovereign zones, aimed at European, Middle East, and Africa markets. For SAP landscapes, this matters less as a real-estate story and more as an information architecture signal: enterprise AI agents will need to act on SAP and non-SAP data while respecting residency, classification, and access policies. According to SAPinsider reporting, 79% of organizations cite data quality, security, and resiliency as requirements, and 73% cite regulatory compliance. That focus explains why OpenText links sovereign cloud infrastructure directly to AI agent deployment rather than treating them as separate projects. The message is clear: without governed, regional information capacity, agentic AI platforms cannot safely scale.
SAP Business AI: Paying Partners for Live Agents
SAP’s €100 million Business AI Partner-Led Adoption Incentive Fund changes how ecosystem projects are financed. Instead of paying for marketing, SAP will pay partners when a customer goes live on an SAP-delivered agent, a Joule Studio application, or a custom AI workflow built on the SAP Business AI Platform. Four funding tiers range from activating a pre-built Joule agent to delivering multi-agent orchestration with workflow applications, tying payouts to concrete AI agent deployment outcomes. This is a response to slow enterprise AI adoption in core SAP scenarios, where most customers remain in identification or experimentation phases. By subsidizing production deployments, SAP is signaling that enterprise AI agents and agentic AI platforms must be operational, not theoretical. It also pushes partners toward native SAP Business AI and Joule Studio, especially as API policies limit third-party agents accessing SAP data from outside the platform.
What the €205 Million Means for Your Architecture
Taken together, OpenText and SAP are committing €205 million to agentic AI platforms and supporting infrastructure, aimed squarely at CIOs, enterprise architects, and systems integrators. The pattern is a shift from AI-enhanced features to AI-native platforms built around enterprise AI agents that can span applications but still respect governance. For architecture teams, this raises three practical questions. First, where will AI agents run—in SAP Business AI, in OpenText’s sovereign cloud infrastructure, or across both via integration? Second, how will identity, entitlements, and audit trails be shared so that agents can act safely across SAP and non-SAP systems? Third, which data products and classifications are required before agents can be turned loose on production workloads? These investments do not remove the need for internal data and process work, but they give IT leaders clearer, production-ready models for AI agent deployment.
Designing Production-Ready Agentic AI Platforms
The early data on adoption shows demand outpacing readiness: SAPinsider research notes that only a minority of organizations are currently implementing AI agents, while many remain in planning or evaluation. That gap is where OpenText’s sovereign cloud positioning and SAP’s partner fund align. Both vendors are building for regulated, data-sensitive environments where data residency, classification, and lineage matter as much as model quality. For practitioners, production-ready AI agent deployment will depend on a few design principles: treat sovereign cloud infrastructure as a jurisdictional control, not a complete security solution; anchor agents to well-defined business workflows in systems of record; and favor platforms that provide clear controls for system boundaries and multi-agent collaboration. Organizations that combine these disciplines with SAP Business AI and OpenText’s agentic AI capabilities will be better placed to move from pilots to durable, auditable agentic AI platforms.
