The launch reflects a broader shift in enterprise software toward AI interfaces that allow users to interact with existing databases and applications through natural-language commands rather than conventional web interfaces or custom code.
NGP VAN Interface adds what the company describes as a fourth interaction layer alongside its web applications, APIs and data products. The system is designed to let compatible AI applications interact with information and functionality already available within NGP VAN platforms while retaining existing access controls.
At the technical level, the offering combines an open-source command-line interface (CLI), an MCP server and a proof-of-concept integration with Anthropic’s Claude. MCP is an open protocol designed to provide AI applications with standardized access to external tools and data sources.
For enterprise software providers, that architecture addresses a central challenge in agentic AI adoption: connecting models to operational systems without simply giving an AI application unrestricted access to sensitive databases.
NGP VAN says its MCP implementation includes purpose-built authentication and permission controls. The platform also introduces an authentication model intended to allow workflows spanning multiple NGP VAN databases to operate through a single user account and command.
That capability could reduce some of the technical friction associated with managing multiple API connections and application environments. It also demonstrates why authentication is becoming a foundational component of agentic software as organizations move from AI-generated content toward systems capable of taking actions.
The natural-language interface is designed to support tasks that would otherwise require users to navigate several applications or write API calls. In practical terms, this changes the interaction model: instead of manually locating information or invoking individual software functions, users can communicate with an AI tool that translates requests into authorized actions.
Security remains a critical consideration in that architecture. Connecting AI agents to operational databases creates new opportunities for automation, but it also introduces risks around permissions, unintended actions, data exposure and auditability. Purpose-built authentication and granular authorization therefore become important components of any production agentic system.
NGP VAN is initially making Interface available through a closed beta. The company says it is working with organizations and technology developers to test natural-language workflows and build agentic capabilities on top of the platform.
The company is also establishing its first Forward Deployed Engineer roles to support integration and implementation. That approach mirrors a growing enterprise software trend in which vendors combine AI infrastructure with hands-on deployment support as organizations move from experimentation toward production use.
Market Landscape
AI interfaces are increasingly becoming an additional layer between users and enterprise software. Platforms from Microsoft, Google, Anthropic and other technology providers are developing agentic architectures that allow AI systems to retrieve information, invoke tools and execute multi-step workflows.
MCP has emerged as one approach for standardizing these connections. Its importance is less about the language model itself and more about establishing a structured mechanism through which AI applications can discover and use authorized external capabilities.
For organizations managing sensitive information, the key question is shifting from whether AI can connect to enterprise data to how that connection can be governed safely.
Strategic Outlook
NGP VAN Interface illustrates how established software platforms are adapting their architectures for agentic AI. The addition of an AI interaction layer could eventually become as important as traditional web applications and APIs.
The larger industry challenge will be balancing convenience with control. As AI agents gain the ability to execute actions rather than simply retrieve information, authentication, authorization, monitoring and human oversight will become increasingly important elements of enterprise AI infrastructure.
Top Insights
- NGP VAN Interface adds an MCP-based AI interaction layer, allowing authorized AI tools to connect with existing enterprise software without replacing conventional APIs.
- Purpose-built authentication and permissions address a central agentic AI challenge: enabling automation while maintaining control over access to sensitive organizational data.
- The single-account model is designed to simplify workflows spanning multiple databases, potentially reducing the operational complexity associated with separate API connections.
- The launch reflects a wider software shift toward natural-language interfaces that allow AI systems to invoke enterprise tools and execute authorized workflows.
Get in touch with our Adtech experts
