Azure Local with Foundry Local can help organisations build new agentic AI solutions by enabling local model execution, support standard agent development patterns and integration with common APIs and SDKs. This allows teams to create intelligent assistants and task-oriented applications that can reason over requests, use tools and operate close to sensitive data. They achieve this whilst also reducing dependency on external connectivity and aligning with security and sovereignty requirements.
How do I host agentic AI on Azure Local?
You might be thinking: ‘Okay, great we have built this ‘thing’, but should we do with it?’
Agentic AI solutions can be hosted on Azure Local by deploying Foundry Local to an Arc-enabled Kubernetes cluster as an Azure Arc extension. Models are then deployed and managed through Kubernetes-native resources. Organisations can then expose secure inference endpoints for agent applications, using API keys or Microsoft Entra ID with TLS protection. This approach supports both CPU and GPU-backed deployments, making it possible to host intelligent assistants and task-oriented agents locally with low-latency performance and stronger operational control.
Why does data sovereignty matter for AI solutions?
Data sovereignty matters because it gives organisations greater confidence over where data resides, who can access it and which legal and regulatory frameworks apply. This is increasingly important as AI solutions begin to process more sensitive operational, customer and business-critical information.
Having a strong grasp on data sovereignty:
- Helps organisations meet regulatory, contractual and internal policy requirements for handling sensitive data.
- Reduces risk by limiting unnecessary data movement across jurisdictions, platforms and third-party services.
- Strengthens trust by showing customers, partners and regulators that data is managed with clear control and accountability.
What does the journey to building an AI solution look like for our organisation?
The journey typically starts by establishing the right local AI platform foundations requiring validated Azure Local hardware with sufficient CPU, memory, storage, networking and accelerator capacity to support the chosen AI models and workload scale. From there, it moves quickly into model deployment and application development. In practice, this means preparing Azure Local for Foundry Local, securing the environment, deploying an initial model and then using standard APIs and SDKs to begin creating agentic AI solutions.
The key stages of the process are:
- Request access to Foundry Local on Azure Local and confirm the target Azure Arc-enabled Kubernetes environment meets Microsoft’s documented prerequisites. Kubernetes orchestrates containerised applications and, with Azure Local and Foundry Local, it provides the on-premises platform layer for deploying, scaling and managing cloud-native workloads and AI services close to where data resides.
- Prepare the platform by configuring Kubernetes, ingress, authentication and certificate management for secure deployment.
- Deploy Foundry Local as an Azure Arc extension and verify that the inference operator is running correctly.
- Review the model catalog, select an appropriate model, and create the first model deployment for local inference.
- Secure access to inference endpoints by using Microsoft Entra ID or API keys together with TLS-enabled connectivity.
- Start application development by integrating with OpenAI-compatible APIs or Microsoft SDKs to create assistants, workflows and task-oriented AI agents.
- Expand from an initial proof of concept into operational agents by refining prompts, adding tools and aligning deployments to CPU or GPU-backed workloads as needed.
BlakYaks, can architect and build the entire Azure Local ecosystem, navigating hardware vendors and helping your organisation to be ready to build and host APIs in as little as four weeks.
It does not need to be complex, allow BlakYaks to demystify the solution.
How else can Azure Local help with data sovereignty?
Beyond local AI inference, Azure Local helps preserve data sovereignty by extending cloud-style operations into customer-managed environments without requiring workloads, data, or operational control to move outside your organisation’s boundaries.
Implementing Azure Local:
- Supports disconnected operations, allowing Azure Local to run without ongoing connectivity to the Azure public cloud in sovereign, regulated and isolated environments.
- Keeps the control plane, identity integrations, monitoring and access control within the local environment, giving organisations greater operational ownership.
- Runs on customer-managed, on-premises infrastructure, helping ensure data residency, governance and compliance requirements can be addressed, without moving sensitive workloads off site.
We already have Azure, can the two co-exist?
Azure Local integrates with Azure through Azure Arc, giving organisations a unified way to manage on-premises resources from Azure while keeping workloads and data local. This integration enables services such as Azure Monitor, Azure Site Recovery, extension-based management and Azure-managed identity support. This allows teams to apply consistent governance, visibility and operational controls across both local and Azure environments.
Microsoft Entra ID and Active Directory integrate with Azure Local and Foundry Local to provide secure identity, access control and policy enforcement across hybrid, on-premises and AI workloads.
What Is the Role of Proact IT UK in an Azure Local Project?
Proact and BlakYaks, can support the design and build of Azure Local and Foundry Local by combining hybrid cloud, Azure platform, Kubernetes, automation and operational delivery expertise. This can help organisations move from platform strategy and architecture to secure implementation, Azure integration and day-two operational readiness.
Proact and BlakYaks’ specialists can support you by delivering:
- Architecture and platform design for Azure Local, Arc-enabled Kubernetes, and Foundry Local deployments.
- Build and automation services using Azure landing zones, Infrastructure as Code, DevSecOps, and cloud native engineering practices.
- Integration with Azure through Azure Arc, governance controls, monitoring, identity, and hybrid operational processes across local and Azure environments.
We are responsible for making all these components work as one coherent system.
What does Proact do in practice?
- Accelerate platform adoption with proven solution design for Azure Local, Arc-enabled Kubernetes and Foundry Local.
- Reduce delivery risk through automated build services using landing zones, Infrastructure as Code, DevSecOps and cloud-native engineering.
- Unlock greater value from hybrid cloud with seamless Azure integration, governance, monitoring, identity and operational alignment.
Proact and BlakYaks help organisations move faster from strategy to deployment, with expert-led design, build and integration services for Azure Local and Foundry Local.
Fast deployment and API readiness
Azure Local with Foundry Local gives organisations a fast route from platform deployment to real AI outcomes. By using Azure Arc-based deployment, Kubernetes-native model management and OpenAI-compatible API patterns, teams can accelerate integration, stand up secure inference services quickly, and begin creating new APIs for agentic AI use cases much earlier in the delivery cycle.
The timeline depends on the complexity of your estate and the required integration, as well as having the right prerequisites in place. A more complex setup will take longer to deliver.
In either case, the Proact programme commits to three things:
- Just weeks to your first AI build → not months to a proof of concept.
- No data movement → data stays where it already lives.
- Ongoing management → Proact operates the stack as a managed service from day one.
How do you know if your organisation is ready to start the AI journey?
Organisations do not need to have solved the AI question before starting the conversation. They need honest answers to two questions that surface where the real work is. These conversations happen at every level.
Why local, structured data accelerates adoption
Organisations are often able to onboard Azure Local and Foundry Local more effectively when the data they want to use is already structured, accessible and held locally. Compared with large volumes of unstructured data spread across multiple repositories, local structured data makes it easier to prioritise use cases, govern access and move faster from deployment into real AI outcomes. This is because:
- Structured, local data is easier to connect to APIs, workflows and agentic AI use cases, helping teams accelerate time to value.
- Keeping data local and organised simplifies governance, access control and compliance, compared with trying to work across fragmented sources and inconsistent formats.
- Starting with structured data creates a stronger foundation for early AI success. While unstructured data distributed across the estate often increases complexity, preparation effort and delivery risk.
Who owns the AI agenda in your organisation?
If the CTO, CISO, or a named data owner is driving the AI programme internally, you can work with a real sponsor. If nobody owns the AI agenda, the first deliverable is a governance and sponsorship framework. This establishes who is responsible for data, who owns the AI outcomes and what success looks like, before any technology is deployed.
Is your organisation ready for Azure Local and Foundry Local?
Your business is ready to start with Azure Local and Foundry Local when the right foundations are in place to turn platform investment into faster AI delivery, stronger data control and earlier business outcomes. The clearest signal is when you can combine technical readiness with well-defined use cases that justify local, secure and sovereign AI services.
Has your organisation got the points below covered?
- You have, or are ready to establish, the Azure Arc-enabled Kubernetes foundation needed to accelerate deployment and shorten time to value.
- You can meet the security and governance prerequisites required to launch local AI services with confidence and maintain stronger control over sensitive data.
- You have identified priority use cases where low latency, data sovereignty, or secure local inference can unlock measurable operational or customer value.
- You are ready to move from planning into action with initial model deployment, API integration and a clear path to early AI outcomes.
BlakYaks can help with readiness workshops, allowing deployment to continue at scale without reduced cadence.
Frequently Asked Questions
What is the difference between Foundry Local on Azure Local and Microsoft Foundry?
Foundry Local on Azure Local brings AI inference into a local, Azure Local environment, while Microsoft Foundry provides a broader platform for building, optimising and governing AI apps and agents at scale.
Does data remain sovereign when using Foundry on Azure Local?
Yes. Microsoft positions Foundry Local on Azure Local as a way to keep data processing on-premises where data is generated, which supports local control, low-latency inference, and sovereignty-sensitive scenarios.
How does it integrate with applications?
Foundry Local on Azure Local supports OpenAI-compatible REST patterns, making it easier to connect applications, inference services and new APIs, by using familiar development approaches.
What do organisations need in order to get started?
During preview, deployment access must be requested. Also, the platform is designed to run on an Azure Arc-enabled Kubernetes environment with the relevant security and ingress prerequisites in place.
Can businesses still integrate with Azure?
Yes. Azure Local integrates with Azure through Azure Arc, which helps unify management, governance, monitoring and extension-based operations across local and Azure environments.
Can teams start building agentic AI solutions quickly?
Yes. Microsoft documentation highlights model deployment, OpenAI-compatible APIs, SDK support and local endpoint patterns, that help teams move quickly from platform setup into AI application and agent development.
Written by…
Carl Harker
Principal Azure Architect
Carl Harker is Principal Azure Architect at Proact IT UK, a specialist IT managed services provider in enterprise storage, hybrid cloud and data infrastructure, since 1994. Proact is a Microsoft solutions partner operating across enterprise and public sector organisations in the UK. Carl started work within the IT industry in 1988, as a trainee Novell Netware Engineer. A pivot to his career came in 2015, supporting early Microsoft Azure systems, moving on to design and implementation from 2020.