Data teams at companies like EY are already running thousands of AI agents across Microsoft Fabric and Copilot to handle tax, audit, and supply chain work with minimal manual input. ZEISS built its next-generation analytics platform on Fabric. PwC uses it to power AI-driven decision workflows at scale.
Microsoft’s 2025 annual report confirmed Fabric as its fastest-growing analytics product ever, now used by 70% of the Fortune 500. Three FabCon 2026 updates pushed that further. Fabric data agents reached general availability, operations agents entered Real-Time Intelligence preview, and Copilot in Fabric gained the ability to modify semantic models in natural language.
In this article, we’ll cover what each update delivers, where it fits, and what enterprises need to do next.
Key Takeaways Fabric data agents reached general availability at FabCon 2026 and now support NL-to-SQL, NL-to-DAX, and NL-to-KQL query generation across lakehouses, warehouses, and semantic models .Operations agents are in active preview within Fabric’s Real-Time Intelligence workload, monitoring live Eventhouse data and triggering automated actions when defined conditions are met. The Power BI Modeling MCP Server lets developers build and modify semantic models using natural language, including bulk renames, measure creation, and DAX validation. Microsoft open-sourced Agent Skills for Fabric on GitHub, with Fabric Local MCP now generally available and Fabric Remote MCP in preview, enabling AI agent integrations across coding tools like GitHub Copilot, Claude, and Cursor. Copilot Studio’s multi-agent orchestration support for Fabric agents is rolling out to general availability, enabling Fabric-grounded agents to run directly inside Microsoft Teams. What Changed at FabCon 2026 for Agents and Copilot FabCon 2026, held in Atlanta in March, produced three categories of updates that directly change how enterprise data analytics teams can work with Fabric . Together, they move the platform from a data orchestration layer into one where agents can read, reason, and act across your entire data estate. The full set of announcements is documented in the Fabric March 2026 feature summary .
The updates do not sit in isolation. Fabric data agents going GA, operations agents entering a live preview, and Copilot in Fabric gaining semantic model modification capabilities are connected. They all share OneLake as a data foundation, Microsoft Purview as a governance layer, and Entra ID as an identity and access framework.
For enterprise teams, the question is no longer whether Fabric can support agentic workloads . The question is how to govern them well. These updates bring new responsibilities around agent permissions, skill governance, and audit logging that teams need to address before scaling deployments.
Fabric Data Agents Are Now Generally Available Fabric data agents moved to general availability at FabCon 2026. These are AI assistants that operate directly over your data assets in OneLake. They translate natural language questions into SQL, DAX, or KQL queries depending on the data source, then return governed, schema-aware answers.
The reach of this GA release is significant. The Copilot in Fabric overview on Microsoft Learn covers the full architecture behind these capabilities. Publishing and sharing data agents within Microsoft Fabric is now generally available, meaning teams can build an agent and deploy it across the organization with full lifecycle management, Git integration, and deployment pipeline support in place.
How Fabric Data Agents Work Each agent connects to one or more data sources: lakehouses, warehouses, Power BI semantic models, KQL databases, mirrored databases, or ontologies. When a user asks a question, the agent pulls the schema using the requester’s credentials, constructs a prompt with that schema, and calls Azure OpenAI to generate the right query.
Three query engines are available depending on the data source. Full schema-based query generation is documented in the Fabric data agent reference on Microsoft Learn. NL-to-SQL handles lakehouses and warehouses. NL-to-DAX handles Power BI semantic models . NL-to-KQL handles KQL databases. The agent picks the right engine automatically based on where your data lives.
Through integration with Microsoft Purview , data agents now support complete auditing, eDiscovery, data lifecycle management , and communications compliance. Every prompt and response generates telemetry that compliance teams can review, making the capability suitable for regulated workflows.
Microsoft Fabric Data Agents: Everything You Need to Know Learn how Microsoft Fabric Data Agents automate data operations, improve governance, and help teams manage analytics workflows more efficiently.
Learn More
What GA Means for Enterprise Deployments General availability means Microsoft carries full support commitments for the feature. For regulated industries, this changes the procurement and risk calculation significantly. Enterprises can now onboard Fabric data agents with confidence that the capability has production-grade SLAs behind it.
Role-based access controls are enforced at the source level. A user querying financial data through a data agent only sees what their Entra ID identity permits. No extra configuration is required at the agent layer. That behavior is what makes data governance in an agentic context tractable for enterprises with complex permissions structures.
Key Use Cases by Function Data agents work best where teams ask the same structured questions repeatedly. Finance teams can query cost allocations and budget performance in natural language. Supply chain teams can ask inventory or logistics questions without waiting for analyst capacity. For HR analytics teams, headcount and attrition queries become self-service instead of ticket-based.
The agent does not replace data analysts . It handles the repeatable, structured query work so analysts can focus on interpretation and the decisions that require judgment. For an industry view of how this plays out, Kanerika’s Microsoft Fabric case studies cover deployments in logistics, manufacturing, and financial services.
Fabric Data Agent Use Cases by Function Function Typical Query Type Data Source Access Control Finance Cost allocation, budget vs actuals Warehouse or semantic model Role-based via Entra Supply Chain Inventory levels, order status Lakehouse or mirrored database Row-level security HR Headcount, attrition rates Semantic model Column-level security Operations KPI dashboards, variance analysis Any Fabric source Purview classification Customer Success Account health, churn signals KQL database or warehouse Object-level security
The access control column matters because Fabric data agents enforce permissions at the source level. Microsoft has published guidance on responsible AI deployment within the platform on the Trusted AI and Fabric blog . Governance policies set in Microsoft Purview flow through to every agent response without additional configuration.
Get Expert Guidance on Agentic AI Deployment with Microsoft Fabric Kanerika brings certified Fabric expertise and hands-on agentic AI experience.
Book a Meeting
Operations Agents in Fabric Real-Time Intelligence (Preview) While data agents answer questions, operations agents watch your data continuously and act when something changes. The Real-Time Intelligence overview on Microsoft Learn explains the full workload architecture that operations agents run within. This is the distinction enterprises need to understand before evaluating either. Operations agents live inside Fabric’s Real-Time Intelligence workload and run against live streaming data, not historical queries.
They monitor live data flowing through Eventhouse and KQL databases, evaluate that data against conditions you define, and take configured actions when those conditions are met. The agent writes its own playbook based on the business goal you set, then runs that playbook autonomously against the incoming data stream .
How Operations Agents Work Setting Up an Operations Agent Getting an operations agent running requires three inputs:
Define your business goal – describe what the agent should monitor in plain language.Specify a data source – currently limited to regular Eventhouse tables within the same workspace.Define the actions – set what the agent does when conditions are triggered, such as sending an alert to Slack, Teams, or Email.How the Agent Runs Once configured, the agent uses Azure OpenAI to generate its own playbook and reason about conditions as they occur. It then runs on a schedule, pulling recent data, evaluating it against the rules, and routing outputs to the configured channel when thresholds are met. All queries are logged in the Eventhouse query insights tab, giving teams a traceable record of what the agent accessed and when.
Before You Deploy: Admin Prerequisites Four requirements must be in place before any deployment: Trial Fabric capacities are not supported. A paid capacity is required. A Fabric administrator must enable the Operations Agent preview toggle in the Admin Portal. Copilot and Azure OpenAI service settings must be enabled at the tenant level. Review pricing in the operations agent billing documentation before going live. What Operations Agents Can and Cannot Do Today As of May 2026, operations agents are in active preview . Several constraints apply before production deployment. Only regular Eventhouse tables are supported. Shortcut tables, functions, and materialized views do not work with the current version.
The feature is not yet available in South Central US and East US regions. For enterprises evaluating this for regulated workflows, the preview status means the feature carries supplemental terms of use rather than full GA SLAs. Compliance and risk teams should factor that into deployment decisions.
Operations Agent Capability Status Capability Status Notes Monitor Eventhouse tables Available Regular tables only KQL database as data source Available Must be in the same workspace Teams / Slack / Email alerting Available Configurable at setup Shortcut table support Not available Regular tables only in current preview Materialized view support Not available Not yet confirmed for GA roadmap Trial capacity support Not available Paid capacity required Compliance-grade audit logging Via Eventhouse query tab Not a standalone audit log format
The table above reflects the current state of the preview. Teams evaluating operations agents for regulated enterprise workflows should verify the latest status via the Microsoft Fabric What’s New page before finalizing deployment plans.
Enterprise Use Cases for Operations Agents Operations agents suit scenarios where human monitoring cannot keep pace with the volume or speed of incoming data.
Manufacturing teams can trigger alerts when production line metrics deviate from tolerance ranges.
Logistics teams can flag delivery exceptions the moment they appear in streaming data, rather than discovering them in the next morning’s report.
Retai l teams can monitor inventory levels across distribution centers and escalate restocking requirements before stockouts occur.
Financial services teams can watch transaction streams for anomalies and route flagged events to compliance reviewers automatically. For each of these scenarios, the agent acts as a first-line response layer that eliminates the latency between an event occurring and a human being informed about it.
For teams already running real-time analytics workloads, operations agents fit naturally into the existing Eventhouse infrastructure without requiring a separate monitoring platform. That reduces both deployment complexity and the surface area for data governance issues.
Copilot Gets New Powers over Semantic Models The Power BI Modeling MCP Server gives Copilot a direct interface to semantic model development. Released as part of the Power BI November 2025 feature updates , it connects GitHub Copilot and compatible AI assistants to Power BI via a local Model Context Protocol server installed as a VS Code extension.
What Natural Language Modeling Means in Practice Through the Modeling MCP Server, developers interact with Power BI semantic models using plain language instead of scripts or manual DAX. Here is what that covers in practice:
Model Building and Updates Create or update tables, columns, measures, and relationships by describing what you need Run bulk operations across a model – renaming hundreds of objects, adding translations, or applying documentation at scale DAX and Validation Generate DAX expressions from natural language descriptions Validate existing measures and query the model to check results, all without switching between separate tools Previously, bulk modifications to a semantic model required scripting in Tabular Editor or writing DAX manually. The MCP server cuts that effort significantly.
Operational Value for Large Power BI Environments For teams managing dozens of Power BI semantic models, the bulk documentation and translation capabilities alone justify the setup. A task that previously took a skilled developer several hours can be completed through natural language prompts in a fraction of the time.
Teams scaling Copilot usage across the organization should also review the Fabric Copilot capacity documentation to understand how billing separates from the capacity running your content.
Report Creation Acceleration The general availability of direct editing for semantic models in the Power BI service, announced at FabCon Vienna in September 2025 , means analysts can modify Live connected and Direct Lake semantic models without downloading a PBIX file. That removes the round-trip between the service and Desktop for model changes.
Together, the MCP server for modeling and expanded in-service editing create a workflow where an analyst can request a new measure in plain language, validate the DAX, add it to the model, and see it reflected in a connected report, all within the same session. For data analytics teams under pressure to deliver reporting faster, that cycle reduction is material.
Open-Source Agent Skills for Fabric Microsoft published the skills-for-fabric repository on GitHub to give developers a curated set of skills and MCP systems for working with Fabric through AI coding tools. The repository supports GitHub Copilot, Claude, Cursor, and Windsurf. It sits within the broader microsoft/skills ecosystem , which covers Azure, AI Foundry, and other Microsoft services.
What the Repository Contains The repository includes skills covering SQL, Spark, Power BI , and KQL workflows, along with MCP configuration for environments that support it. When you clone the repo, it automatically picks up root-level configuration files: CLAUDE.md for Claude Code, .cursorrules for Cursor, .windsurfrules for Windsurf, and AGENTS.md for Codex and similar tools.
Skills give AI assistants guidance and patterns for working with Fabric APIs. MCP servers give them live tool access to data sources. In environments where both are configured, the two layers work together. The AI assistant gains the knowledge to write correct Fabric code and the ability to execute against live environments with the right permissions.
Fabric Local MCP vs Fabric Remote MCP Fabric Local MCP reached general availability at FabCon 2026. It runs on the developer’s machine and gives AI assistants full knowledge of Fabric’s APIs , enabling grounded code generation without connecting to a live environment. It is primarily a developer productivity tool for data engineering and data integration workflows.
Fabric Remote MCP is in preview. It runs as a cloud-hosted server, letting AI agents perform authenticated operations inside a Fabric environment without any local setup. Agents authenticate via Entra ID and operate within existing RBAC boundaries. The server records every tool call in audit logs, which matters for enterprise governance teams evaluating the capability for production use.
Fabric Local MCP vs Fabric Remote MCP Dimension Fabric Local MCP Fabric Remote MCP Status Generally available Preview Deployment Developer machine Cloud-hosted, no local setup Authentication Local access token Entra ID Live operations in Fabric No (guidance and code gen only) Yes (authenticated actions) Audit logging Not included Yes, every tool call recorded Primary use case Code generation for Fabric workloads Agent-driven operations in live environment Compatible AI tools GitHub Copilot, Claude, Cursor, Windsurf Any MCP-compatible client
The Fabric CLI v1.5 GA release is closely related and worth reviewing alongside these MCP tools. The Remote MCP is the more significant enterprise capability because it lets AI agents act in your Fabric environment under your governance policies. The Local MCP is primarily a developer productivity tool that accelerates data engineering and migration work that requires working with Fabric APIs.
Copilot Studio and Fabric: Multi-Agent Orchestration Several multi-agent capabilities in Copilot Studio are rolling out to general availability, giving enterprise teams new options for connecting Fabric-grounded agents to the workflows where business users already operate. The multi-agent updates cover three areas: Fabric integration, Microsoft 365 Agents SDK orchestration, and Agent-to-Agent (A2A) communication.
Fabric Integration in Copilot Studio Copilot Studio-built agents can now work with Fabric data agents to reason over the full data estate without requiring data-intensive engineering work for each query type. An agent built in Copilot Studio can delegate structured data queries to a Fabric data agent, get back governed, schema-aware answers, and incorporate those answers into a conversational response delivered in Teams.
This matters for enterprises where business users work in Teams but analytical data lives in Fabric. The integration closes that gap without forcing users to access Fabric directly, and without requiring engineering effort for each new data-intensive query scenario. For reference, see the Microsoft documentation on consuming Fabric agents in Copilot Studio .
A2A Protocol and Connected Workflows Copilot Studio now supports the Agent-to-Agent (A2A) protocol, which lets Copilot Studio agents communicate directly with third-party agents and specialized internal agents built outside the Microsoft stack. For enterprises deploying custom AI agents across multiple platforms, this is the mechanism that allows those agents to exchange work and context without custom integration code.
Combined with the Microsoft 365 Agents SDK orchestration support, this creates an architecture where Fabric-grounded intelligence can flow across tools from a single interface. Agents handling complex, cross-domain queries can draw on both structured analytics data from Fabric and unstructured knowledge sources simultaneously, returning coherent answers that reflect the full context.
What These Updates Mean for Businesses Across Industries The GA and preview updates to Fabric’s agent and Copilot capabilities change what enterprises can realistically deploy, not just prototype. The practical threshold for each scenario is similar: a governed OneLake data estate, clear role-based access policies, and a human-in-the-loop escalation path for high-stakes decisions.
Industry Use Cases for Fabric Agent and Copilot Enhancements Industry Agent Type Key Scenario Governance Requirement Financial Services Data agent + Copilot Studio Analyst NL queries over secured financial data in Teams Purview classification, row-level security Manufacturing Operations agent Automated monitoring of production line KPIs with real-time alerting Eventhouse telemetry, audit log review Retail Data agent + operations agent Demand signal monitoring and inventory query automation Column-level security, role-based access Healthcare Copilot semantic modeling Semantic model automation for compliance reporting Purview + HIPAA alignment Logistics Operations agent Real-time delivery exception monitoring and alerting Entra ID, full audit trail Insurance Data agent Claims data queries and reconciliation analysis Row-level security, Purview audit
Across industries, the governance architecture is what determines whether deployment is responsible. A useful external framing from Cloud Wars covers how the Copilot Studio and Fabric updates position Microsoft for enterprise multi-agent deployments. Teams already invested in Microsoft Fabric, Microsoft Purview , and Azure Cloud Solutions are best positioned to activate these capabilities quickly because the governance foundation is already in place.
How Kanerika Helps Enterprises Deploy Copilot and Agents in Fabric Moving from an announced capability to a deployed one requires more than technical configuration. It requires understanding how each agent capability fits the data estate you already have, what governance controls need to be in place before users start querying, and how to connect Fabric’s agent layer to the workflows where decisions are made.
Kanerika is a Microsoft Solutions Partner for Data and AI and a Microsoft Fabric Featured Partner. The deployment team holds DP-600 (Fabric Analytics Engineer) and DP-700 (Fabric Data Engineer) certifications, and includes Amit Chandak, Chief Analytics Officer and Microsoft MVP, alongside Fabric Superusers with deep platform knowledge.
As an official FAIAD and RTIAD delivery partner, Kanerika is Microsoft-recognized to conduct certified Fabric training, giving enterprise teams the knowledge transfer they need to operate independently after implementation.
Microsoft Fabric and Purview Expertise Kanerika is also one of the earliest global implementors of Microsoft Fabric , which means the team encountered and resolved the platform’s early-stage edge cases before they became known failure modes. That accumulated experience carries forward into every production deployment.
The team was among the earliest Microsoft Purview implementors globally and holds Microsoft’s Advanced Specializations in Data Warehouse Migration to Azure and Analytics on Azure, validating end-to-end capability across the data modernization path most Fabric deployments require.
Karl: A Production Agentic AI Solution Built Natively on Fabric Karl , Kanerika’s data analytics AI agent, is now generally available as a native Microsoft Fabric workload. Business users get conversational access to lakehouse data in plain English, with no SQL required. Karl was recognized at FabCon 2026 as part of Kanerika’s Fabric innovation track.
Karl is evidence of Kanerika’s ability to build production agentic AI applications that operate within Fabric’s governance model, respect its security boundaries, and deliver measurable value at enterprise scale. That same architectural discipline applies to every client deployment.
Migration Accelerators That Remove the Biggest Barrier to Fabric Adoption Building agentic AI on Fabric requires the data foundation to already be on Fabric. For many enterprises, that means migrating from legacy platforms first. Kanerika’s automated migration accelerators support moves from SSIS, SSAS, Azure Data Factory, Informatica, and Synapse Analytics to Fabric, cutting timelines from months to weeks.
The Azure to Fabric Migration workload is available as a native Fabric workload, featured by Microsoft at Ignite 2025. It automates schema conversions, dependency mapping, connection reconfiguration, and workspace organization. FLIP , Kanerika’s intelligent workflow automation platform, powers migration pipelines end-to-end and is available on the Microsoft Azure Marketplace.
Case Study : Microsoft Fabric Deployed for Real Operational Impact at SSMHSouthern States Material Handling (SSMH), a Toyota Material Handling dealer operating across multiple states, needed a unified data and analytics platform to support faster, more accurate operational decisions across a distributed business.
Kanerika implemented Microsoft Fabric and Power BI to consolidate fragmented data sources, build real-time reporting pipelines, and deliver actionable intelligence to SSMH’s operational and leadership teams. SSMH’s CIO, Delano Gordon, noted: “Kanerika’s flexibility in aligning Microsoft Fabric with our business needs ensures that we are building a system that will drive even better results across our operations.”
85% increase in operational visibility90% data accuracy and KPI reliability100% scalability achieved across the business
Production-Ready AI Agents Deployed Across Client Environments Beyond Karl, Kanerika has deployed six additional purpose-built AI agents in live client environments. DokGPT handles document intelligence via RAG, reducing information retrieval time by 43% and manual review hours by 35% for an investment bank client. Alan handles legal document summarization and clause analysis. Susan handles PII redaction and sensitive data masking. Mike validates quantitative data and financial figures.
Each agent reflects practical experience with governance, integration, and performance challenges that arise in live deployments rather than demo environments. For enterprises assessing their readiness to deploy agentic AI on Microsoft Fabric , Kanerika’s AI Maturity Assessment provides a structured starting point, mapping current data analytics , governance , and AI/ML capabilities against production readiness criteria.
Case Study : Context-Aware AI Agent for Expert RecommendationsThe Client A financial services firm where advisors rely on expert matching to answer client queries. The matching process drew on unstructured research documents, legacy databases, and compliance policies simultaneously.
The Challenges Expert matching was slow and error-prone, driving high mismatch rates and rising support ticket volumes Knowledge was fragmented across sources with no unified retrieval layer The compliance team required every agent response to be traceable to a source, with no hallucinated outputs reaching advisors Different user roles needed different levels of access to sensitive financial data Kanerika’s Solution Kanerika built a context-aware retrieval architecture using RAG, with Microsoft Purview enforcing data classification and access policies at the source level.
Kanerika mapped role-based access controls to the client’s existing Entra ID infrastructure, so the AI agent automatically scoped responses to what each user was permitted to see The team built a human-in-the-loop review layer into the response workflow from day one Full decision path logging gave the compliance team a traceable record of every recommendation and its source documents Business Impact Compliance review shifted from manual review of every output to exception-based review, reducing overhead significantly
80% reduction in mismatch tickets
Zero hallucination incidents post-deployment
Wrapping Up The agents and Copilot enhancements in Fabric announced through early 2026 represent a meaningful shift for enterprise data teams. Fabric data agents are production-ready. Operations agents offer a preview of what continuous, autonomous monitoring can look like at scale. Copilot’s new semantic modeling capabilities reduce the effort required to maintain and extend Power BI data models.
Taken together, these updates give enterprise teams more capability to act on their data estate. The limiting factor for most organizations will not be the technology. It will be the governance, integration, and change management work required to deploy it well. Build the governance foundation now, and your organization will be better positioned to expand agent deployments as more capabilities reach general availability.
FAQs What Are the Latest Copilot and Agent Enhancements in Microsoft Fabric? Microsoft Fabric’s most recent agent and Copilot updates include three major changes. Fabric data agents reached general availability at FabCon 2026, supporting natural language queries across lakehouses, warehouses, and Power BI semantic models. Operations agents entered a live preview in Real-Time Intelligence. The Power BI Modeling MCP Server introduced natural language semantic model development. All three share OneLake and Purview as a common foundation.
What Is the Difference Between Fabric Data Agents and Operations Agents? Fabric data agents answer questions. They translate natural language into SQL, DAX, or KQL queries and return governed answers from your data estate. Operations agents monitor data continuously and act when defined conditions are met. Data agents are generally available. Operations agents are in preview as of May 2026 and currently support only regular Eventhouse tables within the same workspace.
Is Copilot in Fabric Generally Available? Several Copilot in Fabric capabilities are generally available, including report creation assistance in Power BI and semantic model editing in the Power BI service. The Power BI Modeling MCP Server for natural language model development was released in late 2025 as a VS Code extension. Operations agents remain in preview. Each Copilot and agent capability in Fabric has its own availability status, so teams should verify before production deployment via the Microsoft Fabric What’s New page.
What Can the Power BI Modeling MCP Server Do? The Power BI Modeling MCP Server connects AI assistants like GitHub Copilot to Power BI’s semantic model development environment. Through it, developers can create and update tables, columns, measures, and relationships using natural language. They can run bulk operations at scale and validate DAX expressions. It installs as a VS Code extension and works with any compatible AI assistant, giving data analytics teams a faster path from model design to deployment.
What Is Agent Skills for Fabric? Agent Skills for Fabric is an open-source repository on GitHub published by Microsoft. It contains skills and MCP configurations that help AI coding tools like GitHub Copilot, Claude, Cursor, and Windsurf operate over Microsoft Fabric. Skills provide guidance on SQL, Spark, Power BI, and KQL workflows. Fabric Local MCP is generally available and gives AI tools deep knowledge of Fabric APIs for grounded code generation.
How Do Operations Agents Work in Microsoft Fabric Real-Time Intelligence? Operations agents in Fabric Real-Time Intelligence monitor Eventhouse tables continuously. You set a business goal, specify a data source, and define what actions the agent should take when conditions are met. The agent generates its own playbook using Azure OpenAI, evaluates incoming data on a defined schedule, and routes outputs to configured channels like Teams, Slack, or Email when thresholds are triggered. All query activity is logged via the Eventhouse query insights tab, supporting basic governance audit requirements.
Can Fabric Data Agents Integrate with Microsoft Teams? Fabric data agents can be embedded into Microsoft Teams through integration with Copilot Studio. Copilot Studio-built agents can delegate structured data queries to Fabric data agents and return governed, schema-aware answers directly within the Teams interface. This multi-agent orchestration capability is part of Copilot Studio’s 2026 updates that allow business users to query enterprise data without leaving Teams.
What Governance Controls Do Enterprise Teams Need Before Deploying Fabric Agents? Enterprise deployments of Fabric agents require four governance foundations. Role-based access controls must be configured so agents only surface data each user is permitted to see. Microsoft Purview classification should be applied to data sources before agents query them. Decision path logging must capture what each agent accessed and why. And a human-in-the-loop escalation path should be defined for high-stakes recommendations. Kanerika’s data governance practice covers all four foundations for Fabric deployments.
