AI capabilities that companies can use immediately
Microsoft Foundry is Microsoft’s central platform for developing, deploying and managing AI applications and AI agents on Azure. It brings together models, agent capabilities, development tools, governance and monitoring within a single environment. Foundry Tools are available within the ecosystem for ready-made AI capabilities such as natural language processing, translation, computer vision and document analysis.
The services previously bundled under Azure AI Services are now marketed as Foundry Tools and are integrated into the Microsoft Foundry ecosystem. Microsoft Foundry provides access to an extensive catalogue of models, development tools for prompts, RAG architectures and AI agents, as well as operational capabilities for governance and monitoring. The Azure OpenAI Service – the cornerstone of enterprise-grade generative AI – is an integral part of this platform.
The key advantage of Microsoft Foundry and the integrated Foundry Tools over building AI models in-house is the speed at which organisations can get started. Companies do not need to employ their own data scientists, build up training data or operate model infrastructure. They call up AI functions via API and integrate them into their applications. What used to be an AI project lasting several months can be reduced to a matter of weeks with the right services to support development work.
Microsoft Foundry offers organisations extensive options for operating AI applications in accordance with European data protection and compliance requirements. Depending on the model and deployment type, data can, for example, be processed within a selected Azure region or within the EU Data Zone. The architecture, region, network access and permissions can be tailored to the relevant compliance requirements.
Microsoft Foundry has a modular structure: organisations can, for example, start with individual Foundry tools or pre-built models and use these to gradually develop more complex AI applications and agents. This results in AI agents that combine multiple services, access corporate data, control processes independently and operate within controlled, monitored environments. collana develops AI applications using Microsoft Foundry – ranging from simple automation solutions to complete agent architectures for enterprise use.
Azure OpenAI Service provides GPT-4, GPT-4o and other OpenAI models in a secure, enterprise-grade Azure environment. Companies use it to build chatbots and knowledge agents based on their own documents, manuals and databases, automated text generation for reports and communication, and RAG architectures that combine LLM responses with up-to-date company data. collana has developed an AI chat app on Azure OpenAI for Bucher Municipal that makes maintenance data available in seconds.
Relevant for: Wholesale & Foreign Trade, Production & Manufacturing, Healthcare, Logistics & Transport, All knowledge-intensive industries
Azure Document Intelligence (formerly Form Recognizer) automatically extracts structured information from documents: invoices, delivery notes, contracts, forms, ID cards and individual document types are analysed, relevant fields are recognised and output as structured data. Manual data entry, error-prone document processing and time-consuming archiving work are replaced by AI-supported automation.
Azure AI Language provides ready-made NLP functions: sentiment analysis, entity recognition, keyword extraction, text summarisation, question-answer systems and conversational language understanding. Companies can use it to automatically analyse customer feedback, classify service requests, extract relevant information from texts and build conversational interfaces for their applications.
Relevant for: Wholesale & Foreign Trade, E-Commerce, Healthcare, Telecommunications & IT, all companies with high communication volumes
Azure Vision automatically analyses images and videos: object recognition, text recognition (OCR), quality control in manufacturing, facial recognition, scene analysis and custom image classification. In manufacturing, computer vision detects defects on production lines. In retail, it analyses stock levels on shelves. In logistics, it automatically reads labels and number plates.
Azure Machine Learning is the platform for companies that want to train, test and operate their own AI models. AutoML automates model selection, MLOps pipelines control training, versioning and deployment. Data scientists train predictive analytics models for demand forecasting, predictive maintenance and anomaly detection - directly on the company's own data in the Azure environment.
Azure AI Foundry creates AI agents that perform complex tasks independently: Controlling processes, calling up tools, accessing data and coordinating with other agents. collana develops customised AI agents for corporate use: knowledge agents that search internal documentation, process agents that automate approval workflows and service agents that answer customer enquiries independently. In addition, ILAI, collana's own AI platform for chatbots and voice assistants, is available as a ready-to-use solution.
Relevant for: All sectors, in particular Production & Manufacturing, Healthcare, Wholesale & Foreign Trade, Telecommunications & IT
Not because they have larger IT teams. But because they use AI as a tool, not as an experiment. Those who implement Microsoft Foundry aren’t gaining a new technology platform, but concrete automations: documents that read themselves. Service requests that classify themselves. Knowledge agents that find answers in seconds. The following companies already entrust their IT projects to collana.
Many companies know that AI could improve their processes - but the path from idea to productive application seems unclear. Which services fit which use case? How are AI outputs integrated into existing systems? How do you ensure that AI responses are reliable and GDPR-compliant? And how does an initial pilot solution scale to a company-wide application?
collana answers these questions not in theory, but in practice. The AI Discovery Workshop is the structured introduction: specific use cases are identified, evaluated and prioritised in a short space of time. Which process has the greatest potential for automation? Which Azure AI services are suitable? And what does a realistic implementation plan look like that quickly leads to measurable results?
Next comes implementation: collana develops AI applications on Microsoft Foundry that are integrated into existing system landscapes – ERP, CRM, document management systems and web platforms. For companies looking to get started quickly with generative AI, ILAI – collana’s own Azure-based AI assistant – is available as a ready-to-use solution that can be individually configured to suit a company’s data and processes.
AI projects rarely fail because of the technology. They fail because use cases are defined too vaguely, data is not available or not clean enough, or AI outputs are not integrated into operational processes. What is needed is a partner who combines technical AI expertise with an understanding of business processes. This is precisely the approach that collana takes when supporting companies in the introduction and further development of AI applications on Azure AI Services.
collana is a six-time certified Microsoft Solutions Partner and fulfils the requirements in several areas of expertise: Business Applications, Data & AI, Digital & App Innovation, Infrastructure, Security and Modern Work. These certifications are not an end in themselves. They prove that we not only know the Microsoft platform, but that we have mastered it at enterprise level and are continuously developing it further.
What sets collana apart from a pure Azure implementer is its own product experience with AI. ILAI, collana's AI assistant for chatbots and voice assistants, is built on Azure OpenAI and is in productive use with customers. Local LLM solutions for data protection-critical environments, anonymisation solutions for sensitive data and AI developer training for internal teams round off the portfolio. collana knows from its own experience what really works in practice when building AI applications - and what doesn't.
collana accompanies companies not only up to the first AI prototype. The aim is a long-term partnership: from the AI discovery workshop to the development and integration of the first application through to ongoing operation and continuous development. More than 1,000 customers in Germany, Switzerland and other European markets rely on this collaboration.
Not every company knows immediately which AI use case is the right first step. The AI Discovery Workshop provides clarity in a short space of time: concrete use cases, realistic cost estimates and an implementation plan that works with the available budget and existing IT infrastructure.
Microsoft Foundry is Microsoft’s central platform for developing, deploying and managing AI applications and AI agents on Azure. It brings together models, agent capabilities, development tools, and features for monitoring, security and governance.
Foundry Tools are the pre-built AI services within this ecosystem. These include, amongst others, Azure Language, Azure Vision, Speech, Translator and Azure Document Intelligence. They were previously known as Azure AI Services.
ChatGPT is a ready-to-use AI application for end users and businesses. Microsoft Foundry, on the other hand, is a development platform that enables businesses to develop their own AI applications and agents and integrate them into their existing IT infrastructure.
Through Foundry Models, organisations can, amongst other things, deploy OpenAI models within their Azure environment, whilst utilising Azure’s capabilities for identity management, network access, governance, monitoring and data processing. Customer data, prompts and responses from Foundry Models deployed via Azure are not used to train the underlying Foundation Models without explicit consent.
The Foundry Agent Service is Microsoft’s platform service for developing, deploying and scaling AI agents.
Whilst traditional chatbots primarily answer questions or follow predefined dialogues, AI agents can handle multi-step tasks, call up tools and APIs, incorporate data sources and carry out actions in connected systems. For example, knowledge agents can search for internal information, whilst process agents can automate individual steps within a business process.
Microsoft Foundry offers a wide range of features that enable AI applications to be designed in line with European data protection and compliance requirements. Key factors here include the services and models used, the chosen deployment type, the Azure region, and the specific technical and organisational configuration.
Depending on the deployment, data can, for example, be processed regionally or within an EU Data Zone. In addition, features such as Microsoft Entra ID, role-based access control, private networks and other Azure security features are available. Whether a specific solution is GDPR-compliant must therefore always be assessed on the basis of the relevant deployment scenario and architecture.
The costs depend on the model used, the deployment type and the actual usage volume. Many generative AI models are billed on a pay-as-you-go basis, based on the number of input and output tokens processed.
For applications with a predictable high workload, reserved or provisioned capacity models are also available, depending on the model. In addition, there may be costs for further foundry services, data storage, search functions or connected Azure components. A reliable cost estimate can therefore only be provided on the basis of the specific use case and expected usage.
Yes. Microsoft Foundry, Foundry Models and Foundry Tools can be integrated with existing applications and business systems via APIs, SDKs and other Azure integration options.
For example, documents can be processed automatically, service enquiries classified, corporate knowledge made accessible via AI agents, or AI functions integrated directly into ERP, CRM, web and line-of-business applications. The specific integration depends on the existing systems and the particular use case.
A complex agent architecture with multiple data sources and business systems.
In the AI Discovery Workshop, collana therefore begins by assessing the specific use case, the existing data and systems, and the technical requirements. This leads to a realistic assessment of the effort involved, the architecture and possible next steps.