Interview with Horst Urlberger from Microsoft

Horst Urlberger from Microsoft on cloud migration, artificial intelligence and Azure OpenAI Service

"Security and compliance must be the top priority."

As Business Lead Data Modernization at Microsoft, Horst Urlberger opens up new opportunities for companies for the efficient, successful and future-oriented implementation of data & AI solutions from the Microsoft Azure portfolio. In an interview for data! magazine, he explains what exactly it means to modernize data and what fascinating opportunities will open up in the field of artificial intelligence in the coming years.

Horst, what do you do as Business Lead Data Modernization at Microsoft?

In principle, I dealt with market launch topics for our Data & AI services on a daily basis. This includes, for example, developing a local strategy in relation to our Data & AI solution portfolio, marketing activities and promoting partnerships with sales partners, system integrators or technology partners. The sales strategy involves, for example, identifying potential customers, training sales staff and developing sales processes. I also support our sales and technology teams in providing transparency about existing programs that help our customers to use the technologies.

What does "modernize data" mean?

"Modernizing data" refers to the process of updating and optimizing data infrastructures, systems and practices to adapt them to the demands and opportunities of today's digital and increasingly connected world. This enables companies to respond more efficiently to changing business requirements.

What does this look like in detail?

Typically, this involves moving from outdated, inefficient or inflexible data solutions to modern technologies and approaches. By modernizing data, organizations can improve their data management, data processing and data analytics to make informed decisions, drive innovation and gain competitive advantage. It is an important step in realizing the full potential of data and meeting the increasing demands for data volume, quality, speed and variety.

Microsoft has developed a solution to help companies with this.

Microsoft's Intelligent Data Platform is the best integrated and most complete platform for this step-by-step modernization. Using it as a starting point, companies can migrate traditional on-premises databases such as SQL Server or Oracle databases to the cloud, use big data technologies, automate data pipelines and use artificial intelligence and machine learning for data analysis. Overall, the platform helps to build an agile and scalable data infrastructure.

To some ears, this probably sounds like a lot of work at first. In your experience, how do companies react to these changes that they inevitably have to deal with?

Very different, depending on their industry, size and maturity in terms of digital transformation. At management level, there is often a high level of awareness of the importance of data and the need to bring data silos together. Awareness of modernization is correspondingly high. Many recognize that they can use their data more effectively in order to remain competitive or gain advantages. The mood is therefore generally positive, as managers recognize the opportunities and benefits that can arise from data modernization. Nevertheless, depending on the industry, there is still a lack of the right priority for these topics and a lack of willingness to implement them.

‍Andhow is the mood among employees?

The mood among employees is also very positive, especially in the area of data modernization. The ideas within the company are usually very diverse and should be promoted and supported more strongly.

Are they not given enough support?

Not always. I can see that some employees are still being held back by the lack of modern technologies. Yet they are very motivated to use such technologies. When we talk about more traditional digital transformation topics and process optimization, we also still occasionally see reluctance or uncertainty among employees regarding changes in work processes, skills development or the organizational structure. Above all, if management is able to clearly communicate the benefits of data modernization, offer training and support and involve employees in the process, this usually leads to a more positive and committed attitude.

You are responsible for the Basic Program at Microsoft, so you set the framework conditions for making the leap to the cloud. What support can companies expect from you?

Migration and modernization projects are usually divided into five classic phases: Planning, data analysis and data preparation, testing and validation, migration implementation and review and optimization. We use these phases to provide our customers with targeted support. From the first phase, for example, we can support the assessment of existing infrastructure components with tools. Some of these migration tools and services can also be used for the actual migration at a later stage to facilitate and automate the migration process.

Migration is a process that naturally brings with it a whole host of questions. How intensively can you deal with the individual challenges?

By pooling expertise. Microsoft offers the greatest support in the form of consulting services through our certified partners. These services include working with companies to develop a customized migration strategy, providing technical support during the migration process and helping to solve challenges and optimize the cloud environment. In addition, there are usually Azure Credit options for testing and during the migration that we can offer to help with duplicate migration costs, for example.

Many companies are therefore taking their first cautious steps towards digitalization. New technologies in the field of artificial intelligence are already shaking up the market. How are customers picking up the pace to keep up with this rocket science?

A consolidated data platform is needed to really exploit the full potential of AI, as AI models are largely determined by data quality and quantity. However, one of the major advantages of the Microsoft Cloud is the pre-trained cognitive services that are part of Microsoft's AI solution portfolio. One example of this is the Azure OpenAI service. This means that a large number of AI application areas can already be implemented without much effort.

So Azure Open AI Service is a good start for all AI newcomers, so to speak?

You could put it like this. The Azure OpenAI Service is part of the Azure AI Services, and provides access to OpenAI's GPT-x language models such as GPT3, GPT-4, Codex and Embeddings via the Azure platform. Currently, most applications span across tasks such as content creation, summarization, semantic search and natural language to code translation. For example, it can significantly increase end customer satisfaction if the models are used when interacting with customers, for example via improved chatbots or similar. There are also a number of internal fields of application for finding and automating information more quickly, or for summarizing facts more quickly or making them usable. Above all, the use of Azure OpenAI on your own company data usually opens up the most exciting fields of application.

However, specific use cases are required for this.

My clear recommendation is to start by looking for an area of application that can be implemented quickly and therefore demonstrates the benefits clearly and quickly. In parallel, you should then focus on the data platform in order to be able to implement more and more efficient areas of application later on.

Again, two topics at the same time that companies need to take to heart.

Yes, but they are easy to manage. To simplify the first step towards a data platform, we at Microsoft, for example, offer the so-called Data Strategy Assessment, which in turn is carried out by certified partners. This generally provides a good inventory and a clear strategy for the future.

The competition never sleeps. In March, Google announced a serious alternative to ChatGPT in the form of the text robot Bard. Although it is not yet available in EU countries, it is only a matter of time before Bard makes waves here too. How do you differentiate yourselves from the competition?

The most promising areas of application can usually be achieved either in the context of the company's own data or in interaction with end users. In both cases, security and compliance must be paramount. With Azure's industry-leading security and compliance measures, OpenAI's generative AI capabilities can be used with confidence while ensuring data privacy and security.

The responsible use of AI is also very important to Microsoft. Azure OpenAI, for example, has a built-in content filtering system. The system is used to synchronously filter harmful and inappropriate content. The system currently supports the filtering of sexual, hateful, self-harming and violent content. In addition, the system also monitors activity that could indicate some form of abuse so that customers can be warned if someone is abusing their application.

In my personal opinion, Azure OpenAI is much more advanced in terms of "ready to use", which is already confirmed by a large number of international and national reference customers. An excerpt of a few customers who are already so convinced after a short time that they were willing to provide a public reference: Siemens, AT&T, KPMG, EY, Strabag, Carmax, Moveworks as well as a large number of start-ups.

It feels like everyone is currently talking about the possibilities of AI. The role of Microsoft is of course central here. Can you give us an outlook on what we should be looking forward to?

The progress of artificial intelligence is expected to continue to be impressive over the next few years. Machine learning will certainly continue to be a driving force in making AI systems more powerful and precise. New algorithms, techniques and models will be developed to solve complex problems more effectively and make better predictions. The ability of AI systems to understand and process natural language, images and videos will be further improved. In the near future, I assume that audio, image and video input can be used much more effectively, meaning that multimodal models will drive innovation potential.  

And how do you think companies will react to these innovations?

Companies will increasingly understand how to use AI technologies effectively in different areas. We will see a broader application of AI in areas such as healthcare, finance, energy, manufacturing, transportation and other industries. AI will help provide more efficient processes, personalized services and innovative solutions.

As AI progresses, the discussion about ethical aspects and responsibility will also intensify. Questions of data protection, transparency, fairness and responsibility of AI systems will become increasingly important. Regulation and framework conditions could be further developed in the coming years to promote the responsible use of AI. Overall, this will increasingly lead to AI becoming even more suitable for mass use and seamlessly integrated into our lives.

Thank you very much, Horst!

The interview first appeared in the first issue of data! You can find all issues and articles here.

‍