Book a consultation

Microsoft Foundry: AI agents & apps for business.

Microsoft Foundry (formerly Azure AI Foundry) is Microsoft’s platform for your own AI agents and AI applications. NAVAX connects it to Business Central, Dynamics 365 and Microsoft 365.

Employee with a headset at a workstation with several monitors
Three people talking at a table with a laptop in front of a window front
Trusted by companies including
Microsoft Solutions PartnerISO 27001 zertifiziert (Information Security Management)
At a glance

What is Microsoft Foundry?

Microsoft Foundry is Microsoft’s platform in Azure for building, deploying and governing AI agents and AI applications. It brings models, agents, tools and company data under one management layer, with monitoring, evaluation and access rights. Until it was renamed, it was called Azure AI Foundry.

What
Consulting, pilot and implementation of AI agents and AI apps on Microsoft Foundry, formerly Azure AI Foundry.
For whom
Companies in Austria and Germany that want to move AI from the test phase into everyday work.
Result
Prioritized use cases, an assessed business case, a working pilot and a plan for scaling.
Who
NAVAX, Microsoft Solutions Partner for Business Applications, with Dynamics 365 and Business Central implementations.
Why NAVAX
Microsoft technology, industry know-how, process understanding and AI expertise from one source. NAVAX already runs multi-agent systems in production itself.
Challenges

From AI experiment to platform: three typical hurdles.

Many companies are already experimenting with AI. It only adds value once it runs securely, at scale and directly in business processes. These are the three statements NAVAX hears most often.

  1. Individual pilots, nothing in daily use

    Today

    “We have a few AI pilots, but none of them is used day to day.” Every attempt starts from scratch again.

  2. The AI doesn’t know our data

    Today

    General AI tools answer with knowledge from the internet, not with your master data, documents and records.

  3. IT and data protection slow things down

    Today

    “IT and data protection are holding us back because nobody has the full picture.” Nobody knows which agent accesses which data.

In practice

Several agents, one result.

NAVAX already runs multi-agent systems in production, for example the NAVAX PMO Buddy for project management.

  1. In-house multi-agent system
    NAVAX PMO Buddy
  2. Task

    Analyze status, risks, resources and reporting across many projects together.

  3. Setup

    An orchestrator coordinates specialized agents, each handling one subtask.

  4. Effect

    Supports analyses, reports and decisions across projects and portfolios.

For whom

Is Microsoft Foundry right for you?

Foundry is a platform for development teams. If more of the right-hand column applies, another path is often better, and NAVAX will tell you so in the first conversation.

Tap the statements that apply to you.

A good fit if …

Another route is better if …

Your result
Nothing selected yet.

Select the statements above that apply to you. This is where you will see whether a conversation is worth it.

Services

NAVAX services for Azure AI Foundry, now Microsoft Foundry.

Whether you use Azure AI Foundry, Azure OpenAI or no AI platform at all today, NAVAX meets you where you are.

  • Two colleagues talking in front of a large digital whiteboard, one pointing at the screen

    Microsoft Foundry strategy workshop

    Together we work out prioritized AI use cases and their business value. We review your data and system landscape, clarify governance requirements and outline the target architecture and roadmap. You leave the workshop with a clear basis for decisions.

    More on AI consulting
    WorkshopRoadmap
  • Employee at a desk behind two monitors

    Implementing a pilot

    NAVAX implements the most important use case as a pilot. You see early on what the agent achieves with your real data and decide based on results rather than assumptions. Typical use cases with Dynamics 365 and Business Central:

    • Check and improve master data quality automatically
    • Document and knowledge assistants for records, contracts and manuals
    • Support and service agents that triage and answer requests
    • Support for quotes and in sales
    Foundry Agent ServiceDynamics 365Business Central
  • Three colleagues standing in front of a green moss wall

    Scaling successfully

    If the pilot proves itself, NAVAX moves it into regular operations and extends it to other departments: sales, service, purchasing, finance, HR and project management. For complex tasks, NAVAX combines several specialized agents that work together.

    Multi-agentMonitoringGovernance
Process

How a Microsoft Foundry project works with NAVAX.

Four stages: find potential, assess the business case, implement a pilot, scale. That way, you only invest more once the value is proven.

  1. 1

    Find potential

    We review your processes and data and collect use cases where AI noticeably saves time or prevents errors. Lead time: one to two weeks.

    What you have afterwards

    A list of possible use cases from your processes.

  2. 2

    Assess the business case

    For the most promising use cases, we assess value, effort and risks. If needed, a first prototype shows whether the idea holds up technically. Lead time: up to one week.

    What you have afterwards

    An assessed business case and the decision on which use case goes into the pilot.

  3. 3

    Implement a pilot

    The selected agent runs with real data for a first group of users. We measure the results and refine. Lead time: three to four weeks.

    What you have afterwards

    A working pilot with measured results.

  4. 4

    Scale successfully

    We move the agent into regular operations, establish monitoring and governance and transfer the pattern to other areas. Lead time: four to twelve weeks, depending on complexity and scope.

    What you have afterwards

    An agent in regular operations and a pattern for further areas.

NAVAX in numbers

We don’t just implement. We stay.

NAVAX builds its own software and has been running it for years. That is why we know what a system has to withstand in its tenth year.

30
Years in business
330
People
1,000+
Projects
8 locations in Austria and Germany
Microsoft Solutions PartnerISO 27001 certified (Information Security Management)
Microsoft Solutions Partner for Business Applications. ISO 27001 certified.
FAQ

Frequently asked questions about Microsoft Foundry and Azure AI Foundry.

What people often ask before a first conversation. If your question is missing, ask NAVAX directly.

What is Microsoft Foundry?

Microsoft Foundry is Microsoft’s enterprise platform in Azure for building, deploying and governing AI agents and AI applications. It brings models, agents, tools and company data under one management layer, including monitoring, evaluation and access rights. The model catalog offers models from Microsoft, OpenAI and many other providers.

Is Azure AI Foundry the same as Microsoft Foundry?

Yes. Microsoft renamed Azure AI Foundry, previously Azure AI Studio, to Microsoft Foundry. People searching for “Azure Foundry” or “Microsoft AI Foundry” usually mean this platform as well.

How much does Microsoft Foundry cost?

Microsoft Foundry is billed based on usage through your Azure subscription. Costs arise mainly from model usage and from tools and data connections. NAVAX estimates realistic costs for your use case together with you in the strategy workshop.

What is the difference between Azure and Foundry?

Azure is Microsoft’s cloud platform; Microsoft Foundry is a service within Azure built specifically for AI agents and AI applications. To use Foundry in the cloud, you therefore need an Azure subscription. Foundry is managed as its own resource with shared roles, network and policy rules.

What are Microsoft Foundry agents?

Microsoft Foundry agents are AI agents you create and run in Foundry Agent Service. An agent combines a model with instructions and tools so that it understands requests and completes tasks on its own. Options include prompt agents without custom code and hosted agents with your own code, for example based on Microsoft Agent Framework or LangGraph.

Microsoft Foundry vs. Copilot Studio: which fits whom?

Copilot Studio is the low-code tool business departments use to build agents for Microsoft 365 and Teams quickly. Microsoft Foundry is the platform for development teams when you want to choose models freely, add your own code, evaluate agents systematically or orchestrate several agents. In many projects, the two complement each other.

Microsoft Foundry vs. Azure OpenAI: what is the difference?

Azure OpenAI provides OpenAI models in Azure; Microsoft Foundry is the broader platform around them. In Foundry, you use OpenAI models alongside many others and add agents, tools, evaluation and governance. According to Microsoft, existing Azure OpenAI resources can be upgraded to a Foundry resource while keeping the endpoint and API keys.

What is Foundry Local?

Foundry Local is Microsoft’s option for running AI models directly on the device. The data does not leave the device, the application also works offline, and you do not need an Azure subscription for it. This suits applications with sensitive data or without a stable internet connection, for example.

Is Microsoft Foundry related to Palantir Foundry?

No. Palantir Foundry is a data platform from the company Palantir; Microsoft Foundry is Microsoft’s AI platform in Azure. The two products only share a name.

Can Microsoft Foundry connect to Business Central and Dynamics 365?

Yes. Agents on Microsoft Foundry can use data from Business Central, Dynamics 365 and Microsoft 365 if the connection is set up properly. This is exactly where NAVAX focuses, because NAVAX knows the data models and processes of these systems from its ERP and CRM projects.

How GDPR-compliant are agents built with Microsoft Foundry?

That depends on the individual use case: which data an agent uses, which model is deployed and who has access. Microsoft processes your data as a processor, and with deployments in the EU Data Zone, data and processing generally stay in the EU; individual add-on services may differ. Under the EU AI Act, the obligations depend on the risk class of the application.

Start with your first Foundry use case.

In the strategy workshop, NAVAX clarifies which use case pays off first for you, which data the agent needs and what a pilot looks like. Afterward, you know whether and how to start with Microsoft Foundry.

Read on

You might also be interested in.

Related solutions around AI agents and data, and articles from the NAVAX blog.