MicroStrategy pushes semantic layer for agents

Strategy, the company formerly known as MicroStrategy, spent this week arguing that agentic AI needs a semantic layer before it can be trusted in production. On its events page, the company listed a May 21 webinar titled “Why Data Governance must work with the Semantic Layer,” with a second session scheduled for May 26. The pitch was that governed metrics, reusable business logic and centralized controls can give AI systems a consistent source of truth across analytics and automation. Hush Security, which operates as Hush Protocol, made a parallel case on May 21 around agent runtime security. Its website says AI agents are moving from assistants to autonomous actors and argues that existing security stacks were built for human users rather than software agents. The overlap between the two messages is operational: one side is selling governed data and business semantics, the other runtime controls and execution security. (strategy.com) ### What exactly is Strategy selling to agent builders? Strategy’s current product materials describe Mosaic as a “universal semantic layer” that standardizes business definitions, metrics and governed access to data. The company says centralized policies can be declared once and then enforced across BI dashboards, AI agents and custom applications, rather than being recreated in separate tools. (hushprotocol.com) Strategy’s May 2026 product notes also show where agents fit in that stack. The release notes say Mosaic model objects can now be displayed while viewing dashboards, the Universal Semantic Layer, agents, cards and Auto Narrative, tying agent interfaces directly to the same governed modeling layer used elsewhere in the platform. (www2.microstrategy.com) ### Why does the semantic layer matter for agent behavior? Strategy’s own framing is that fragmented data sources create “multiple versions of the truth,” which in turn raises hallucination, audit and compliance risks for enterprise AI. Its semantic-layer page says the fix is to move governance out of siloed applications and into a shared layer that translates technical data into standardized business terms. (www2.microstrategy.com) That same page links the semantic layer to explainability. Strategy says a governed semantic layer gives AI agents a trusted source of metrics and KPIs and makes AI decisions “auditable and grounded in business logic.” It also ties that layer to Model Context Protocol, or MCP, saying AI orchestrators can connect through a dedicated MCP server that validates permissions and pulls authorized data in real time. (mstrategybi.com) ### Where does Hush fit in? Hush Protocol’s public site focuses on securing AI agents as they take actions, not just as they answer questions. The company says agents are beginning to make purchases and control systems, and that those behaviors require a different security model from legacy cloud approvals and human click-based workflows. That puts Hush on the execution side of the stack. (mstrategybi.com) Strategy is describing how an agent should retrieve governed context and business logic; Hush is describing how the same agent should be controlled and observed when it starts using tools or interacting with systems. That reading is an inference from the two companies’ published materials, not a joint statement. (hushprotocol.com) ### What does “observable” mean in this setup? Strategy’s language centers on audit visibility, policy consistency and governed access. Its May 2026 release notes introduced Mosaic Sentinel as a governance intelligence platform with “real-time risk alerts, comprehensive audit visibility, and clear usage insights” across the semantic layer. (mstrategybi.com) Those details matter because they describe an agent stack that can be inspected at more than one stage: what data definitions were used, what permissions applied, and what usage was recorded. Hush’s security pitch extends that chain into runtime behavior by focusing on autonomous agents acting inside systems. ### What happens next? Strategy’s events page shows the next public marker on May 26, 2026, when it plans to rerun the semantic-layer governance webinar for a U.S. audience. (www2.microstrategy.com) Hush Protocol’s site continues to market its AI agent security platform, while Strategy’s May 2026 documentation points customers to Mosaic, Mosaic Sentinel and agent-facing features already listed in the current release. (strategy.com) (mstrategybi.com)