--- title: "Google Agentic Data Cloud vs Context Layer: Who Owns the Meaning?" url: "https://atlan.com/know/ai-agent/gcp/google-agentic-data-cloud-vs-context-layer/" description: "Compare Google Agentic Data Cloud with a neutral context layer: current feature status, business meaning, agent access, portability, and coexistence." author: "Karthik Pasupathy" author_role: "Contributing Writer, AI Context & Agents" published: "2026-10-27" updated: "2026-10-27T00:00:00.000Z" --- > Atlan is hosting Context Conference, bringing together the leaders and builders at the frontier of giving AI the context it needs to understand their business. It runs online on October 28, 2026, from 11:00 AM to 2:00 PM ET. Atlan co-founder Prukalpa Sankar opens and closes the day. Leaders from AstraZeneca, BNY and Verizon share why they invest in context and what they get from it. Registrants get early access to The AI Context Gap, a new study from MIT Technology Review Insights. Register: https://atlan.com/context-conference/ ## How does Google Agentic Data Cloud compare with a neutral context layer? Google Agentic Data Cloud brings technical and business context together in Knowledge Catalog, with a Looker integration that adds semantic definitions and lineage. A neutral context layer performs similar work in a separately governed system designed to serve Google and other connected platforms. The choice depends on where the enterprise wants to govern the authoritative copy of each shared definition: within Google's architecture, or in the neutral layer. Four practical differences - **Platform role**: Google combines context with the services that store, process, analyze, and act on data - **Context inputs**: Knowledge Catalog draws on Google and supported external systems; a neutral layer uses its own connectors - **Governance authority**: approved definitions sit in Google's catalog, or in an independent system that serves every platform - **Portability**: Google supports external agents and metadata export; a neutral layer reduces dependence on one metadata model ## Key takeaways - Google Agentic Data Cloud combines data, context, and agent tools across Google and supported external systems. - A neutral layer earns its place when one definition must stay identical across Google and at least one other platform. - The two can work together by syncing supported context while keeping data, execution and meaning in the right systems. - Google Agentic Data Cloud can be enough when coverage, ownership, agent access, cost and portability meet enterprise needs. --- The comparison comes down to four practical differences: * **Platform role:** Google combines context with the services that store, process, analyze, and act on data. A neutral layer governs and serves context across the data platforms and agents already in use. * **Context inputs:** Knowledge Catalog brings together technical metadata, business definitions, quality signals, relationships, unstructured content, and Looker semantics from Google and supported external systems. A neutral layer derives context from both Google and non-Google systems via its connectors. * **Governance authority:** A Google-native design keeps approved definitions within Google's catalog and semantic architecture. A neutral design keeps the authoritative definitions in an independent system that serves Google and other platforms. * **Portability:** Google supports external agents and metadata export, so it is not a closed system. A neutral layer is designed to reduce dependence on a single platform's metadata model, although actual portability still depends on connector coverage and how definitions translate across systems. --- ## What does Google mean by Agentic Data Cloud? Agentic Data Cloud is Google's architecture for bringing data, context, and agent-driven work together. Google Cloud [introduced it on April 22, 2026](https://cloud.google.com/blog/products/data-analytics/whats-new-in-the-agentic-data-cloud), combining capabilities across its existing services. An enterprise can use Google to manage context within this architecture and Atlan, the Context Layer for AI, as the [enterprise context layer](https://atlan.com/know/what-is-the-enterprise-context-layer/) for definitions that must remain consistent across Google and other platforms. Atlan can bring supported context from Google and the rest of the enterprise into one place, where teams approve shared definitions and send supported updates back to Google. Agentic Data Cloud brings the following Google services and capabilities together: * **BigQuery and operational databases:** Stores, processes, and queries business data across analytical and application workloads. * **Knowledge Catalog:** Aggregates context, business definitions and relationships so agents can discover relevant assets. * **Borderless Lakehouse:** Connects Google query engines to supported data and catalogs across clouds. * **Data agents and Conversational Analytics:** Helps practitioners build pipelines, perform analysis and work with data through natural language. * **MCP servers and Data Agent Kit:** Exposes supported Google data tools to agent applications and coding environments. [Google Knowledge Catalog](https://atlan.com/know/ai-agent/gcp/google-knowledge-catalog/) was formerly Dataplex Universal Catalog. Google's [release notes](https://docs.cloud.google.com/knowledge-catalog/docs/release-notes) confirm that API and IAM names remain unchanged; its [pricing documentation](https://cloud.google.com/products/knowledge-catalog/pricing) says SKU names still reference Dataplex. The bundle therefore covers several layers of the stack. Agentic Data Cloud is an architecture rather than a single product release, so availability must be checked on a component-by-component basis. --- ## Which components of Google's Agentic Data Cloud are GA and which are in Preview? Agentic Data Cloud combines established Google Cloud products with newer capabilities that remain in Preview. Availability must be checked at the feature and integration level because a GA product can still contain Preview connectors, semantic features or agent surfaces. According to Google's public announcements, documentation and release notes, this is the status picture as of October 27, 2026: | Capability | Current status | Google's public evidence | | :---- | :---- | :---- | | Knowledge Catalog metadata aggregation, high-precision semantic search and data products | GA, with some source integrations still in Preview | [Knowledge Catalog announcement on April 22, 2026](https://cloud.google.com/blog/products/data-analytics/introducing-the-google-cloud-knowledge-catalog); [data products GA on May 25, 2026](https://docs.cloud.google.com/knowledge-catalog/docs/release-notes) | | Knowledge Catalog connectivity to Salesforce Data360, ServiceNow and Workday | Preview in the April 22, 2026 announcement; no later GA release located as of October 27, 2026 | Knowledge Catalog announcement on April 22, 2026 | | Looker (Google Cloud core) metadata cataloging, including LookML, and Knowledge Catalog lineage for Looker | GA: metadata cataloging on October 21, 2026, lineage on October 26, 2026 | [Looker release notes on October 21, 2026](https://docs.cloud.google.com/looker/docs/release-notes); [Knowledge Catalog release notes on October 26, 2026](https://docs.cloud.google.com/knowledge-catalog/docs/release-notes); [LookML in Knowledge Catalog GA announcement](https://cloud.google.com/blog/products/business-intelligence/lookml-metadata-indexing-in-knowledge-catalog-is-now-ga) | | Automated context curation, verified queries and semantic guardrails | Preview in the April 22, 2026 announcement; no later GA release located as of October 27, 2026 | Knowledge Catalog announcement on April 22, 2026 | | Self-learning context enrichment in Knowledge Catalog | Preview, launched October 22, 2026 | [Self-learning context enrichment announcement](https://cloud.google.com/blog/products/data-analytics/knowledge-catalog-self-learning-context-enrichment-for-ai-agents) | | Knowledge Studio in Knowledge Catalog | Preview, launched October 22, 2026 | [Self-learning context enrichment announcement](https://cloud.google.com/blog/products/data-analytics/knowledge-catalog-self-learning-context-enrichment-for-ai-agents) | | BigQuery measures | Announced in Preview on April 22, 2026; no GA announcement found | [Agentic Data Cloud announcement, April 22, 2026](https://cloud.google.com/blog/products/data-analytics/whats-new-in-the-agentic-data-cloud); [BigQuery measures documentation](https://docs.cloud.google.com/bigquery/docs/graph-measures) | | Borderless Lakehouse federation with AWS Glue, Databricks Unity Catalog, Snowflake Horizon and Workday Data Lake | Preview in the Lakehouse release notes | [Lakehouse release notes](https://docs.cloud.google.com/lakehouse/docs/release-notes); [Workday Data Lake documentation](https://docs.cloud.google.com/lakehouse/docs/set-up-cross-cloud-connection-workday) | | SAP Business Data Cloud integration, including querying SAP data and publishing data products back to SAP | GA | [Lakehouse release notes on July 20, 2026](https://docs.cloud.google.com/lakehouse/docs/release-notes) | | BigQuery Graph | GA; automatic ingestion of Graph metadata into Knowledge Catalog remains Preview | [BigQuery release notes on August 31 and September 14, 2026](https://docs.cloud.google.com/bigquery/docs/release-notes) | | BigQuery remote MCP server | GA | [BigQuery release notes on April 20, 2026](https://docs.cloud.google.com/bigquery/docs/release-notes); [BigQuery announcement on April 22, 2026](https://cloud.google.com/blog/products/data-analytics/unveiling-new-bigquery-capabilities-for-the-agentic-era) | | Conversational Analytics API for BigQuery and Looker | GA | [Conversational Analytics API release notes on June 23, 2026](https://docs.cloud.google.com/gemini/data-agents/conversational-analytics-api/release-notes) | | Conversational Analytics in BigQuery | GA | [BigQuery release notes on June 23, 2026](https://docs.cloud.google.com/bigquery/docs/release-notes) | | Data Engineering Agent and Data Science Agent | GA | [Data Engineering Agent announcement on April 22, 2026](https://cloud.google.com/blog/products/data-analytics/exploring-the-data-engineering-agent-in-bigquery/); [Colab Enterprise release notes on May 26, 2026](https://docs.cloud.google.com/colab/docs/release-notes) | | Knowledge-Catalog-powered Deep Research Agent in Gemini Enterprise | Preview | [Google data-agent announcement](https://cloud.google.com/blog/products/data-analytics/new-data-agents-across-the-agentic-data-cloud/) | | Data Agent Kit | GA | [GA announcement on September 30, 2026](https://cloud.google.com/blog/topics/developers-practitioners/data-agent-kit-is-now-ga-bring-google-data-cloud-to-any-coding-agent) | For an architecture review, evaluate the exact path each workload uses. A GA product can still depend on a Preview connector, semantic feature or agent surface. --- ## How far does Google now reach across clouds and agent vendors? BigQuery Omni already lets BigQuery query data in Amazon S3 and Azure Blob Storage without copying it. Google's 2026 releases go further on two boundaries that once made [platform-native architectures](https://atlan.com/know/why-platform-native-context-layers-fail/) easy to dismiss: catalog federation across clouds and which coding agent a team uses. Google's [borderless Lakehouse](https://docs.cloud.google.com/lakehouse/docs/release-notes) lets BigQuery and Spark query supported data across clouds without first moving it into Google Cloud. Its catalog federation connects with [AWS Glue](https://atlan.com/know/ai-agent/aws/amazon-glue-data-catalog-vs-cross-cloud-governance-layer/), [Databricks Unity Catalog](https://atlan.com/know/ai-agent/semantic-layer/unity-catalog-vs-cross-platform-semantic-layer/), and [Snowflake Horizon](https://atlan.com/know/snowflake/snowflake-horizon-context-and-atlan-context-layer/). These federation paths, along with the Workday Data Lake connection, remain in Preview as of October 27, 2026. The SAP Business Data Cloud integration is an important exception to a purely inward flow. It lets BigQuery query SAP data and supports publishing Iceberg tables and Knowledge Catalog data products back to SAP. This integration is GA. Google has also expanded beyond its own agent interfaces. [Data Agent Kit](https://cloud.google.com/blog/topics/developers-practitioners/data-agent-kit-is-now-ga-bring-google-data-cloud-to-any-coding-agent) brings Google Cloud data tools and skills into Claude Code, Codex, Cursor, and other supported development environments. These agents can work with Google services using the enterprise's existing Google Cloud permissions. This makes the comparison less about connectivity. Google can access supported data outside Google Cloud and make its data tools available to agents beyond Gemini. The case for a neutral context layer instead depends on whether the enterprise needs a single, independently governed source for shared definitions across Google and the rest of its stack. The remaining architectural question is where the authoritative copy of each shared definition is stored and how that definition is propagated to every connected platform and agent. --- ## Who authors, controls, and pays for the context? [Business owners author the meaning](https://atlan.com/know/context-layer-ownership-data-vs-ai-teams/). The architecture determines where the approved version is stored, how changes are controlled, which agents can retrieve it, and which services incur costs. Consider net revenue. Finance decides whether to include credits, refunds, taxes, and regional exceptions. In a Google-centered architecture, the approved definition can be maintained through Google's catalog and semantic services. With a neutral context layer, the authoritative definition is maintained independently and mapped into Google and other connected platforms. Whichever design the enterprise chooses, one version must be authoritative. Control covers who can approve a change, which version replaces the previous one, and how agents recognize the approved definition. Dion Hinchcliffe, CIO Practice Lead at The Futurum Group, summarized the problem in [InfoWorld's April 2026 coverage](https://www.infoworld.com/article/4162737/google-pitches-agentic-data-cloud-to-help-enterprises-turn-data-into-context-for-ai-agents.html): "The hardest AI problem is inconsistent meaning." Giving several agents access to the same tables does not resolve it. Choosing where context resides is not only an architectural decision. It also determines which services the enterprise must configure, operate, and pay for. Knowledge Catalog API calls, including catalog search, are free, while metadata storage, processing, and jobs that run through services such as BigQuery, Dataflow, or Cloud Storage may incur charges. Data lineage, data quality and data profiling run on the premium processing tier, billed in DCU-hours under SKUs that still carry the Dataplex name. A neutral design adds the cost of the context platform while the underlying data-service costs remain. [Google describes Data Agent Kit](https://cloud.google.com/products/data-agent-kit) as available at no charge to Google Cloud users. The Google Cloud services it invokes may still generate usage charges. Google also lets customers [export Knowledge Catalog metadata](https://docs.cloud.google.com/knowledge-catalog/docs/export-metadata) as JSONL files in Cloud Storage. [Context lock-in](https://atlan.com/know/ai-agent/context-layer/single-stack-lock-in-vs-neutral-context-layer/) is therefore less about downloading records and more about reconstructing their behavior elsewhere. A practical [context portability](https://atlan.com/know/ai-agent/context-portability/) test is whether the exported definition, relationships, and version still produce the intended interpretation in another system. Jim Hare, VP Analyst at Gartner, warned in the same InfoWorld article: "Exiting Google-managed semantics, Gemini agents, or BigQuery abstractions may be harder than migrating data alone." This is an analyst warning about reimplementation effort, not evidence that Google prevents exports. For [data sovereignty for AI agents](https://atlan.com/know/ai-agent/data-sovereignty-for-ai-agents/), identify the owner, authoritative system, downstream copies, and services that process each definition. Once those responsibilities are clear, the next task is ensuring every agent retrieves the same approved version. --- ## How can Gemini, Google Agent Development Kit, Claude and in-house agents use the same meaning? Enterprise AI agents can use the same business meaning when each workflow is configured to retrieve the same approved definition, version, and scope before acting. The agents do not need to use the same model or interface, and their answers do not need to be worded identically. Google's [BigQuery MCP documentation](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) provides one example. The remote Model Context Protocol (MCP) server supports Antigravity, ChatGPT, Claude, and custom applications. It lets these clients query data, retrieve metadata, and list resources through authenticated Google Cloud identities. To keep the meaning consistent, the enterprise must configure each agent workflow with an authenticated path to the approved context: | Agent surface | Configuration to establish | | :---- | :---- | | [Gemini Enterprise](https://atlan.com/know/ai-agent/ai-agent-applications/google-agentspace/) | Route the request through a configured data agent or context source that uses the approved definition | | Agents built with [Google Agent Development Kit](https://atlan.com/know/ai-agent/ai-agent-applications/google-adk-vs-langgraph-vs-autogen/) | Connect the tools, APIs, or MCP servers that retrieve the definition and preserve its version and scope | | Claude or ChatGPT | Configure the MCP endpoint, authenticated identity, permitted operations, and approved context source | | In-house agents on [Vertex AI or Microsoft Foundry](https://atlan.com/know/ai-agent/ai-agent-applications/azure-ai-foundry-vs-google-vertex-ai/) | Configure the selected APIs or MCP clients to preserve the definition identifier, version, and access controls | This is where the distinction between an [MCP gateway and a context layer](https://atlan.com/know/ai-agent/mcp-gateway-vs-context-layer/) matters. [MCP can deliver business context](https://atlan.com/know/mcp-delivers-business-context/) to an agent, but the connection does not decide which definition is authoritative. The [A2A protocol](https://atlan.com/know/google-a2a-protocol/), created by Google and now a Linux Foundation project, supports interaction between agents, but it does not replace business ownership of the meaning they exchange. When configuring [MCP servers for enterprise data](https://atlan.com/know/ai-agent/how-to-build-mcp-servers-for-enterprise-data/), test the same business question across every client. Confirm that each one retrieves the same definition version, reporting period, and scope while respecting its permissions. The wording may differ, but the business interpretation should remain consistent. --- ## What does a neutral context layer hold beyond Google Agentic Data Cloud? A [neutral context layer](https://atlan.com/know/ai-readiness/context-layer-101/) holds the enterprise context that must remain [consistent across cloud platforms](https://atlan.com/know/ai-agent/context-layer/multi-cloud-context-layer/), data tools and agent clients. Google Agentic Data Cloud can manage many of the same definitions, relationships and policies within its architecture. A neutral layer provides the enterprise with a single place to approve and maintain the context that Google and other connected platforms should share. This includes: * **[Approved definitions and metrics](https://atlan.com/know/ai-agent/how-to-manage-kpi-definitions-and-a-business-glossary-in-a-data-catalog/):** The business meaning and calculation rules that every platform and agent should use. * **Scope and exceptions:** Where a definition applies, including relevant regions, business units, reporting periods and use cases. * **Ownership and [version history](https://atlan.com/know/ai-agent/context-versioning-for-ai-agents/):** Who approved the definition, when it changed and which version is current. * **Cross-platform relationships:** How business concepts connect to data assets, dashboards, models, policies and processes across the estate. * **Rules for use:** Which agents and teams can access the context and how they should handle [conflicting or outdated information](https://atlan.com/know/context-drift-ai-agents/). Technical metadata can remain in [native catalogs](https://atlan.com/know/ai-agent/gcp/google-knowledge-catalog-vs-neutral-catalog-bigquery/), while a [semantic layer](https://atlan.com/know/semantic-layer/) provides executable calculation logic. The [catalog, semantic layer, and context layer](https://atlan.com/know/ai-agent/semantic-layer/context-layer-vs-data-catalog-vs-semantic-layer/) can coexist, with the neutral context layer connecting their inputs to business decisions governed across the enterprise. The practical test is whether Google's architecture can govern and serve this shared context everywhere the enterprise needs it. --- ## When is Agentic Data Cloud enough on its own? Agentic Data Cloud can be enough when Google is the primary environment for the data, semantic services, catalog, and agent workflows in scope. A BigQuery-and-Looker estate centered on Gemini is the clearest fit, although non-Google agents can also connect through supported Google interfaces. To determine whether Google's native architecture is sufficient, ask the following questions: * **Coverage:** Are the required sources, connectors, and features available, or is the team willing to accept the risks of Preview capabilities? * **Authority:** Can business owners approve and maintain definitions within Google's catalog and semantic services? * **Agent access:** Does every required agent have a tested, authenticated path to the same approved context? * **Operations and cost:** Can the team manage changes, monitor usage, and budget for the underlying services? * **Portability:** Do Google's exports and the work required to rebuild context elsewhere meet future migration needs? If these conditions hold, a separate context layer may duplicate capabilities the enterprise can operate within Google. Where they do not, especially when shared meaning must remain authoritative across several platforms, a [build vs. buy vs. bundle assessment](https://atlan.com/know/ai-agent/context-layer/context-layer-tco-build-vs-buy-vs-bundle/) should determine how Google and a neutral layer divide responsibility. --- ## How can Atlan and Google work together in a GCP-heavy estate? Google can remain the source for native service metadata, permissions, and execution, while Atlan provides a shared control point for business context that must remain consistent across Google and the rest of the estate. Google's Knowledge Catalog announcement lists Atlan among its supported partner catalogs. Together, the two platforms can support the following workflow: 1. **Bring Google context into Atlan:** Atlan's [Knowledge Catalog connector](https://docs.atlan.com/apps/connectors/data-governance/knowledge-catalog/references/what-does-atlan-crawl-from-knowledge-catalog) ingests supported Aspects, data-quality results, and profiling information for corresponding BigQuery assets. 2. **Connect context from the wider estate:** Teams can add source-specific connectors to build [data lineage across platforms](https://atlan.com/know/ai-readiness/ai-ready-data-lineage/), pipelines and BI tools. The Knowledge Catalog connector does not provide lineage by itself, so coverage depends on the other connected systems. 3. **Review and approve shared meaning:** [Context Engineering Studio](https://docs.atlan.com/product/capabilities/governance/context-engineering-studio/concepts/what-is-context-engineering-studio) helps teams assemble, test, improve and deploy context. [Context Agents](https://docs.atlan.com/product/capabilities/governance/context-agents-studio) can help draft descriptions, glossary terms, and other context, while domain owners decide what is approved. 4. **Return supported updates to Google:** With reverse sync enabled, custom Aspect field values edited in Atlan are written back to Knowledge Catalog for BigQuery tables, views, materialized views, columns, schemas and routines. Google's system aspects are never written back, and Aspect schemas remain read-only in Atlan, so teams must define which system owns each field. 5. **Serve approved context to agents:** Atlan's [remote MCP server](https://docs.atlan.com/product/capabilities/atlan-ai/how-tos/remote-mcp-overview) can deliver context to Gemini CLI, [Google Agent Development Kit](https://atlan.com/know/ai-agent/ai-agent-applications/what-is-google-adk/), Claude, ChatGPT, and custom clients. Google's services and each agent runtime continue to enforce their own execution permissions. Atlan's [Context Lakehouse](https://atlan.com/context-lakehouse/) adds an Iceberg-native context store that can live in customer-owned Google Cloud Storage and make context available through MCP, A2A, SQL and APIs. This lets BigQuery remain central to data processing while approved business context can be used beyond Google's architecture. Atlan says this context is stored in Apache Iceberg and is readable by any Iceberg-compatible engine, and the same portability test applies to it: an exported Atlan definition should still produce the intended interpretation without Atlan. Know which path each Google agent actually reads. Looker (Google Cloud core) metadata cataloging, including LookML, went GA in Knowledge Catalog on October 21, 2026, and Knowledge Catalog lineage for Looker followed on October 26, so in many GCP estates the executable formula for net revenue [stays in LookML](https://atlan.com/know/ai-agent/semantic-layer/lookml-vs-universal-semantic-layer/). Atlan's documented paths into Google are custom Aspect values through reverse sync and its MCP server for agents configured to call it. Atlan's docs list Context Engineering Studio deployment targets as Snowflake Cortex Analyst, Databricks Genie, dbt and Claude over MCP, with no Looker or BigQuery target, so decide which system holds the executable formula and which holds ownership, scope and version. Deployment details still matter. Context Engineering Studio requires account-team enablement, and the availability of Context Agent templates varies. Teams should confirm the required connector coverage, synchronization fields, and product availability before assigning responsibilities. Used together, Google provides native data services and an execution environment, while Atlan carries [governed business context](https://atlan.com/know/agent-context-layer/) across the wider estate. --- ## What does shared context look like in practice? Two Atlan customers show what a shared context layer looks like at scale. CME Group brought more than 18 million assets and 1,300+ glossary terms into a [single context layer](https://atlan.com/context-lakehouse/) in its first year with Atlan, helping teams reuse the same context across the business. DigiKey uses Atlan as a shared source of context across its enterprise systems, including the operational systems where business processes run. Neither story is a Google-specific agent deployment. They show catalog-scale adoption; consistent agent behavior still depends on how each workflow retrieves and applies that context. --- ## Real stories from real customers: Shared context at enterprise scale "Context is the differentiator. Atlan gave our teams the shared vocabulary and lineage to move from reactive data management to proactive AI enablement." - Kiran Panja, Managing Director, Cloud and Data Engineering, CME Group "Atlan is our context operating system to cover every type of context in every system including our operational systems." - Sridher Arumugham, Chief Data Analytics Officer, DigiKey --- ## How should you choose the right architecture? Start with one business definition that already matters. Ask its owner to approve the scope, connect two agents, and check which version each receives. Then change the definition and test whether the update reaches both. Use Google's native architecture when it can govern and deliver the approved definition to all required systems and agents. Add an [independent context layer](https://atlan.com/know/how-to-implement-enterprise-context-layer-for-ai/) when teams must repeatedly reconcile or recreate that definition across platforms. --- ## FAQs about Google Agentic Data Cloud vs a context layer ### 1. What is Google Agentic Data Cloud, and what does it include? Google Agentic Data Cloud brings Google's data, context and agent capabilities into one architecture. It includes BigQuery, Knowledge Catalog, borderless Lakehouse, data agents, MCP interfaces, and Data Agent Kit. Each component has its own role and release status. ### 2. Which Agentic Data Cloud components are GA and which are in Preview? As of October 27, 2026, BigQuery Graph, Data Agent Kit, Google's data engineering and science agents, Knowledge Catalog metadata aggregation, search and data products, and Looker (Google Cloud core) cataloging and lineage in Knowledge Catalog are GA. Several federation paths, catalog integrations, and semantic features remain in Preview. Check the release status of every component before implementing the workflow. ### 3. Is the borderless (cross-cloud) Lakehouse the same as a neutral context layer? No. Borderless Lakehouse provides access to supported data across clouds. A neutral context layer maintains shared business meaning independently and makes it available across connected platforms and agents. ### 4. Can non-Google agents (Claude, ChatGPT, in-house) use Agentic Data Cloud context? Yes. Google's BigQuery MCP server supports Claude, ChatGPT, Antigravity, and custom applications. Data Agent Kit also brings Google Cloud data tools into supported coding environments, subject to authentication and permissions. ### 5. Does Knowledge Catalog replace a third-party context layer? It can meet the requirement when the necessary sources, definitions, and agents fit within Google's supported architecture. A neutral layer is useful when shared meaning must span platforms under a single approval process. The decision depends on where the authoritative definitions need to reside. ### 6. Where do business definitions live if both Google and a neutral layer define them? Each shared definition should have one authoritative source. Other platforms can consume synchronized copies, while native service observations remain authoritative in their source systems. Clear ownership and conflict rules prevent local changes from replacing approved definitions. ### 7. How hard is it to move context out of Google-managed semantics later? Google supports metadata exports, so the records themselves can be moved. The larger effort is rebuilding platform-specific calculations, relationships, permissions, and integrations. Test an exported definition in another runtime before treating file portability as functional portability. ### 8. When is Agentic Data Cloud enough on its own? It can be enough when Google's supported services cover the required sources, definitions and agent workflows. Business owners must also be able to approve changes, manage costs, and accept the migration effort. Using non-Google agents alone does not require a separate context layer. --- ## Sources 1. [Google Cloud: What's new in the Agentic Data Cloud, April 22, 2026](https://cloud.google.com/blog/products/data-analytics/whats-new-in-the-agentic-data-cloud) 2. [Google Cloud: Introducing Knowledge Catalog, April 22, 2026](https://cloud.google.com/blog/products/data-analytics/introducing-the-google-cloud-knowledge-catalog) 3. [Google Cloud: New BigQuery capabilities, April 22, 2026](https://cloud.google.com/blog/products/data-analytics/unveiling-new-bigquery-capabilities-for-the-agentic-era) 4. [Google Cloud: Data Agent Kit GA, September 30, 2026](https://cloud.google.com/blog/topics/developers-practitioners/data-agent-kit-is-now-ga-bring-google-data-cloud-to-any-coding-agent) 5. [Google Cloud: Data Agent Kit product page](https://cloud.google.com/products/data-agent-kit) 6. [Google Cloud: Lakehouse release notes](https://docs.cloud.google.com/lakehouse/docs/release-notes) 7. [Google Cloud: Knowledge Catalog release notes](https://docs.cloud.google.com/knowledge-catalog/docs/release-notes) 8. [Google Cloud: BigQuery release notes](https://docs.cloud.google.com/bigquery/docs/release-notes) 9. [Google Cloud: Work with measures](https://docs.cloud.google.com/bigquery/docs/graph-measures) 10. [Google Cloud: LookML in Knowledge Catalog is now GA, October 26, 2026](https://cloud.google.com/blog/products/business-intelligence/lookml-metadata-indexing-in-knowledge-catalog-is-now-ga) 11. [Google Cloud: Looker release notes](https://docs.cloud.google.com/looker/docs/release-notes) 12. [Google Cloud: Self-learning context enrichment in Knowledge Catalog, October 22, 2026](https://cloud.google.com/blog/products/data-analytics/knowledge-catalog-self-learning-context-enrichment-for-ai-agents) 13. [Google Cloud: Workday Data Lake cross-cloud connection](https://docs.cloud.google.com/lakehouse/docs/set-up-cross-cloud-connection-workday) 14. [Google Cloud: Conversational Analytics API release notes](https://docs.cloud.google.com/gemini/data-agents/conversational-analytics-api/release-notes) 15. [Google Cloud: Data Engineering Agent](https://cloud.google.com/blog/products/data-analytics/exploring-the-data-engineering-agent-in-bigquery/) 16. [Google Cloud: Colab Enterprise release notes](https://docs.cloud.google.com/colab/docs/release-notes) 17. [Google Cloud: New data agents across the Agentic Data Cloud](https://cloud.google.com/blog/products/data-analytics/new-data-agents-across-the-agentic-data-cloud/) 18. [Google Cloud: Use the BigQuery MCP server](https://docs.cloud.google.com/bigquery/docs/use-bigquery-mcp) 19. [Google Cloud: Export Knowledge Catalog metadata](https://docs.cloud.google.com/knowledge-catalog/docs/export-metadata) 20. [Google Cloud: Knowledge Catalog pricing](https://cloud.google.com/products/knowledge-catalog/pricing) 21. [InfoWorld: Agentic Data Cloud and analyst perspectives, April 23, 2026](https://www.infoworld.com/article/4162737/google-pitches-agentic-data-cloud-to-help-enterprises-turn-data-into-context-for-ai-agents.html)