LoginSign Up
← Back to all sparks
Holistics logo

Holistics

ANALYTICS
Velocity7.5

Self-service BI and data analytics platform

www.holistics.io ↗

Holistics connects to warehouse-native semantic layers, shifting from semantic owner to governed exploration layer.

bi-analyticssemantic-layerwarehouse-nativeai-governancecode-first-bidata-modeling
◆Current state
Holistics is a code-first BI platform built around AML (Analytics Modeling Language). The current release cadence shows two parallel threads: strengthening its code-first fundamentals (Markdown as first-class analytics objects, explicit interaction syntax for Canvas dashboards) and adding governance for AI access to data (per-user MCP controls, AI query tracing that identifies which surface triggered a warehouse query). The UX is also being refreshed — a new universal sidebar navigation shipped alongside dark/light mode theming. Shipping cadence is high; almost every week brings a user-visible change.
◆Where it's heading
The Snowflake and Databricks semantic layer integration is the most strategically significant move in the current cycle. Rather than requiring customers to migrate their semantic definitions into Holistics, the product now consumes warehouse-native semantics directly. This repositions Holistics from a tool that owns the semantic model to a visualization and embedding layer that works with wherever semantics are already defined. Paired with AI governance features (MCP access controls, AI trace), the trajectory points toward Holistics becoming the exploration and AI-safe access layer sitting on top of the data warehouse — not competing with it.
◆Prediction
If Snowflake and Databricks semantic integration gains traction, the next logical expansion is to BigQuery and Redshift semantic layers. The AI governance thread — per-user MCP controls and query tracing — will likely deepen as enterprise customers demand auditable records of what AI agents are reading and running.

◆Recent moves

  1. 1d ago

    Holistics integrates Snowflake and Databricks semantic layers

    ⚡ SPARK

    Holistics now syncs governed definitions directly from Snowflake semantic views and Databricks metric views, letting teams build dashboards and embeds without maintaining a duplicate semantic layer inside Holistics. This is the most directional move in the current cycle — the product is explicitly stepping back from owning the semantic model when the warehouse already does.

    View source ↗
  2. 8d ago

    🏷 Exclude Endorse & Archive tags when duplicating objects

    Duplicating an object in Holistics no longer carries over its Endorse or Archive status — a small but correct behavior fix that prevents accidentally endorsing a copy before it's been validated. No trajectory relevance; pure housekeeping.

    View source ↗
  3. 10d ago

    🧭 New navigation

    Holistics replaced its fragmented app navigation with a single universal sidebar covering dashboards, data, and tools. The black header is gone. This is a meaningful surface-area consolidation that simplifies onboarding, though it doesn't change what the product can do — it changes how discoverable those capabilities are.

    View source ↗
  4. 12d ago

    🔍 Trace AI-triggered queries and jobs

    AI-triggered queries are now identifiable in the Jobs API and warehouse query comments, exposing which surface initiated a query and how the caller authenticated. This lands alongside per-user MCP access controls, forming a coherent AI governance layer for teams that need auditability around what their AI agents are running against the warehouse.

    View source ↗
  5. 12d ago

    Dynamic light and dark mode dashboard theming

    Dashboard theming now supports dynamic light and dark mode, following the earlier dark mode launch. For embedded analytics customers, this matters: dashboards can match the host app's theme without separate embed configurations.

    View source ↗
  6. 16d ago

    Explicit interaction syntax for Canvas Dashboard

    Canvas dashboards now use an explicit `explicit_interactions` list to declare every active filter connection, removing ambiguity about which filters affect which visualizations. This is a code-first correctness improvement — the kind that matters for teams managing dashboards as code, not a change end-users notice directly.

    View source ↗