PostHog Semantic Layer Gives Analytics Queries a Shared Business Vocabulary.

PostHog has released Semantic Layer, a governed dictionary that attaches shared definitions to company data. The beta is designed to make analytics queries more reliable by giving both people and AI agents a consistent account of what business terms and metrics mean.

The company announced the release on LinkedIn on August 20, following an official blog post published a day earlier by Thiago Rocha Salvatore. PostHog describes the feature as a catalog of definitions that sits on top of existing data and tells users and agents what to trust.

That may sound like data infrastructure plumbing, but it addresses a visible problem. Asking a database for “active customers,” “revenue,” or “conversion rate” is only useful if everyone agrees on how those concepts are calculated. Otherwise, two analysts can produce different answers from the same underlying records, while an AI assistant may choose an unintended table, property, or interpretation.

Semantic Layer creates a place for those meanings to live alongside the data. Instead of relying on context scattered across dashboards, documentation, query logic, and employees’ memories, an organization can establish governed definitions that analytics systems can reference when answering questions.

For a small software company, that could mean ensuring that “weekly active user” always follows the organization’s accepted definition. A subscription business could similarly distinguish gross revenue from recognized revenue, or define whether test accounts and refunded transactions should be excluded from a metric. PostHog has not published a complete list of supported definition types in the supplied announcement, so the precise scope of the beta should not be assumed beyond the company’s description of a governed catalog.

The AI angle is important. Natural-language analytics can make data accessible to more employees, but a fluent answer is not necessarily a correct one. Giving an agent an approved semantic framework can reduce ambiguity before it generates a query or explanation. It also creates a clearer path for teams to inspect the business assumptions behind an answer rather than treating the output as a black box.

PostHog’s release comes as the company is putting more analytics work into automated and conversational interfaces. Its recent changes include native charts in Slack analytics answers and AI-assisted investigation features. A shared semantic layer can become foundational in that environment because every additional interface needs to interpret the same data consistently.

The immediate value will depend less on the existence of the layer than on the quality of the definitions entered into it. Teams considering the beta should first identify metrics that routinely cause disagreement, then document their owners, filters, time windows, and source data. A semantic layer can distribute agreed meaning, but it cannot settle unresolved business decisions on its own.

Useful evaluation questions include:

• Whether definitions are easy for non-specialists to inspect and understand
• How changes to a definition are governed and communicated
• Which PostHog query surfaces and agents currently use the definitions
• How existing dashboards and historical analyses behave after definitions change

PostHog says users can join the beta through the information provided in its blog post. The supplied sources do not specify pricing, plan requirements, or a date for general availability. That makes Semantic Layer an early feature to test against real, disputed metrics rather than something teams should immediately treat as a finished source of truth.

What is PostHog Semantic Layer?

PostHog Semantic Layer is a governed dictionary of definitions attached to an organization’s data. It is intended to help people and AI agents use trusted, consistent meanings when querying that data.

Who is the Semantic Layer for?

It is relevant to teams that use PostHog data across analysts, operators, product staff, and AI-powered analytics workflows. It may be particularly useful where different people or queries currently produce conflicting interpretations of the same business metric.

Is PostHog Semantic Layer generally available?

No general availability date was provided in the supplied sources. PostHog describes Semantic Layer as a beta and directs interested users to its official blog post for information about joining.