Hoval manages an extensive portfolio of heating and indoor climate systems with numerous product variants.
Data alone does not create added value. And AI is only as good as the data, processes and systems to which it has access. collana therefore links data, business intelligence and artificial intelligence along a shared value chain.
Develop a vision and identify opportunities.
Consolidate data from all relevant sources.
Transparent key figures for better decisions.
Automate processes and deploy AI agents.
Data residency, network isolation and access control.
Microsoft Fabric, Foundry, Power Bl and Azure.
Many AI initiatives start with the model. The crucial questions often lie one level below:
Where is the required data located? Is it accessible? Are the definitions consistent? Who is authorised to access it? And how does an AI demo become an application that works reliably with ERP, CRM or other operational systems?
The same applies to business intelligence. A dashboard can visualise data, but it cannot resolve conflicting data sources or unclear business logic.
Understanding objectives, processes and requirements
Collect and integrate relevant data
Consolidating, modelling and providing data
Analysing data and enabling decision-making
Leveraging corporate knowledge and automating processes
01
We provide clarity on data readiness, target architecture, governance and the key areas for action. This results in a robust roadmap for your data and AI initiatives. If the strategic direction has not yet been determined, our Data Strategy Assessment for businesses offers a structured starting point.
02
We consolidate data from ERP, CRM, e-commerce, production and other data sources to create a scalable foundation for reporting, analytics and AI. Microsoft Fabric enables data integration, data engineering and analytics to be combined on a single platform.* If the target architecture is already clear, a data sprint puts the data foundation into practice.
03
We integrate data sources, standardise key performance indicators and develop reports, dashboards and self-service BI solutions for management and business units. We primarily use Microsoft Power BI and TARGIT, ranging from targeted reporting to an international BI platform.
04
We develop AI solutions that harness corporate knowledge, automate processes and intelligently enhance existing business systems. Microsoft Foundry enables the development and management of AI applications and agents. For specific use cases, our AI solutions lead directly to implementation.
Search through corporate knowledge, develop knowledge agents or integrate generative AI securely into existing processes.
We develop scalable solutions based on the Microsoft platform and integrate them with your data and applications.
A modern data and AI architecture does not consist of a single product. Different technologies perform different tasks.
Microsoft Fabric brings together data engineering, data warehousing, analytics and business intelligence on a single, integrated data platform. Predictive models can be trained and run on the same platform, for example to forecast utilisation, demand and capacity.
In regulated sectors, decisions are made before the first use case is implemented: where data is processed, who can access it, and whether this can be verified. We address these questions at the design stage, not after the fact.
Data and AI services do not need to be accessible via the public internet. We connect Fabric and the associated Azure services to your virtual network via private endpoints (Azure Private Link). Access to Fabric is then not via the public internet. We use managed private endpoints to enable Fabric to access your data sources.
Integrate, store, transform and model data, and make it available for analytics or other applications.
Develop models and agents, integrate corporate knowledge, and deploy and manage AI applications.
Depending on the use case, both
Platforms should form part of the same architecture.
Data and AI projects do not have to start with a transformation programme lasting several years.
A sensible starting point often involves a specific problem: a manual report, a lack of transparency across several companies, corporate knowledge that is difficult to access, or a process with significant potential for automation.
This can gradually lead to the development of a scalable architecture:
Understanding the data landscape, objectives and use cases.
Integrate data, ensure quality and define responsibilities.
This ensures that, rather than isolated data or AI projects, reusable capabilities are developed within the organisation.
Hoval manages an extensive portfolio of heating and indoor climate systems with numerous product variants.
At Pilatus, computer-based training is a key component of the training programme. The audio for the...
At Bucher Municipal, technical maintenance knowledge is made accessible via an AI application. The ...
Altano International is consolidating operational and financial key figures using a centralised BI solution...
At the van hire company Carl & Carla, the AI agent ILAI handles recurring processes relating to...
An international BI platform based on TARGIT and Microsoft technologies was developed for Beckhoff ...
Data and AI projects do not end with the data platform. The real added value is created when data and AI are integrated with ERP, CRM, e-commerce, specialist applications and real-world business processes.
It is precisely these systems that teams within the collana Group implement and operate.
The collana Group holds the Microsoft Solutions Partner designations for Data & AI and Business Applications. This enables us to combine Microsoft-validated expertise in data and AI with knowledge of the business applications and processes in which corporate data is generated.
Our teams cover a range of roles: strategic consultants, data architects, data engineers, analytics engineers, AI developers, software engineers and ERP consultants. This enables us to develop a vision and then support its technical implementation.
Not every business needs a central data platform straight away. We can start with a clearly defined reporting, data or AI use case and then expand the architecture in a controlled manner. In the same way, we support businesses that are already facing the challenge of standardising data and analytics across the organisation or internationally.
Data & AI describes the integration of data management, analytics and artificial intelligence. This includes, for example, data strategy, data integration, data platforms, business intelligence, machine learning, generative AI and AI agents. The aim is not to deploy as many technologies as possible, but to make corporate data usable in such a way that it leads to better decisions, more efficient processes and new digital applications.
Many AI applications require business data in order to deliver relevant and reliable results.
If data is inaccessible, inconsistent or poorly documented, even a powerful AI system is difficult to use productively.
Data quality, access rights, governance and a suitable data architecture are therefore key foundations for scalable AI.
Business Intelligence primarily prepares company data for reporting, analysis and management.
Artificial intelligence can also generate content, recognise patterns, make predictions or automate tasks and processes.
Both areas frequently draw on the same database
Power BI is the reporting and analytics layer within Microsoft Fabric, where it runs as an integrated workload on the same data foundation. At the same time, Power BI can still be licensed independently and can be used without Fabric. It was Power BI Premium per Capacity that was discontinued, not Power BI itself. Which option makes sense depends on whether, in addition to reporting, data engineering, data warehousing and machine learning are also to run on the same platform.
We determine this together as part of the architecture. Depending on your requirements, we process data in a specified Azure region, in a private cloud environment, or entirely on-premises within your own infrastructure. For data that must not leave the corporate network, we run language models on-premises. The allocation depends on your data classification.
Yes. We operate the services in isolated virtual networks and connect them via Private Link, so that data traffic does not pass over the public internet. Identities and authorisations are managed via Entra ID, and access is controlled on a role-based basis.
No.
A clearly defined AI use case can often be launched using just a few carefully selected data sources.
However, as the number of applications, departments or data sources increases, a centralised and well-governed database becomes increasingly important for scalability, quality and operations
No.
The Microsoft platform can integrate data from a wide range of Microsoft and non-Microsoft systems.
These include, for example, ERP and CRM systems, e-commerce platforms, databases, APIs, marketing systems, production data and bespoke business applications.
The best way to start depends on your starting point.
If there is no shared vision, a data strategy assessment may be useful. For specific data issues, a data sprint is a good option. For reporting requirements, a BI project can be launched straight away. Where the potential of AI is still unclear, relevant use cases are first evaluated and prioritised.
No. Fabric is an option worth considering if you want data engineering, data warehousing, analytics and reporting to run on a single platform. Alternatively, you can build a data platform using individual Azure components, tailored to your requirements. We’ll determine which approach is best during the architecture phase.
