Business Operational Context for Agentic AI

Entity-centric data products

Entity-centric data products form the core runtime layer of the platform. Each data product exposes governed, entity-scoped context backed by Micro-DBs, as described in Section 03. In the architecture, data products sit between enterprise data sources and the consumers that require business context, including data agents, applications, and analytics tools.

Data agents

Data agents provide the AI-facing control layer in the architecture. They receive requests from AI agents, determine the relevant entity and data product, apply runtime policies, and choose the appropriate access path, such as MCP, API, or direct Micro-DB query.

Data agents are built using K2view’s no-code Data Agent Builder, which supports data retrieval, summarization, tool selection, guardrails, testing, and validation.

Data product access methods

K2view data products expose several governed access methods, depending on the consumer and use case.

MCP provides a standardized protocol for AI systems and agent frameworks. A data product can expose an MCP server that makes the entity data model available as resources, predefined operations available as tools, and optional prompts available to guide agent interactions.

APIs allow enterprise applications, integration services, and data agents to interact

9