AI & Robotics

AI Factory: Overcoming Boundaries with Artificial Intelligence

Artificial intelligence (AI) offers unique potential. It not only helps existing business processes to become more efficient and faster. Above all, its targeted use makes it possible to open up fields of activity that were previously unused - and thus to create valuable new benefits. This is precisely where ERGO's AI Factory comes in.

https://www.ergo.com/aifactory

One thing is certain: AI is increasingly providing impetus in the digitally oriented world of ERGO and thus supports the declared goal of becoming the leading digital insurer by 2025. This also includes a digital "data culture" which helps to exploit ERGO's enormous treasure trove of data and consistently exploit the potential of AI. The new "AI Factory" plays a key role here as the technological basis for using AI in business processes. In this way, AI contributes to a sustainable increase in ERGO's business success. "The AI Factory can be seen as a kind of central system that introduces 'intelligence' into other core application systems and thus into ERGO's business processes", explains Felix Wenzel, Head of Data Engineering at ERGO Digital Ventures AG and technical product owner of the AI Factory.

Felix Wenzel, product owner of the AI Factory

An example: In order to identify even more opportunities for additional business via cross-selling or up-selling and thus generate leads, a much deeper content analysis - for example of customer conversations - is necessary than today. The decisive factor here is, among other things, the consideration of the entire customer communication, which, however, requires the use of cognitive abilities. This quality has so far been attributed exclusively to humans and not to machines or systems specially developed for this purpose. However, the effort required to have humans systematically perform this task would far exceed the benefits. This is precisely where AI, and thus the AI Factory, comes into play. ERGO can already point to promising successes in this context. For example, a significant improvement in the success rate of so-called autonomous outbound calls, i.e. the active establishment of telephone contacts on the basis of AI-supported call analyses, has been demonstrated.

The goal: concrete benefits

The AI Factory is firmly grounded in application practice. It is not concerned with the scientific new or further development of artificial neural networks, but concentrates on the application of artificial intelligence in the form of algorithms that address specific insurance-relevant issues. The goal is concrete benefit.

From the very beginning, the aim was to design an architectural landscape oriented towards open source IT standards, the quality of which lies in particular in breaking down previous boundaries and exploiting new opportunities: cloud-based, fully compatible, flexible, transparent, professional and future-oriented. A high standard that has already proven itself in reality and has enormous potential for realisable digital future scenarios. By providing such an infrastructure, many of ERGO's business units can already achieve goals in the short and medium term which would be a long way off without the use of such a central platform. With this approach, ERGO is already a leader in the insurance industry in the field of AI.

"With AI, a powerful new digital technology is emerging. Fast payouts for immediate claims, increasing sales or restoring customer confidence - AI will play an important role in terms of how ERGO operates in the future", emphasises Robert Meisner in his role as technical product owner for the AI Factory at ERGO Technology & Services Management AG (ET&S). "The AI Factory is an integrated IT platform developed by the ET&S Tech Hub Warsaw. It provides software developers, data scientists and data engineers with the opportunity to optimise their machine learning workflows, scale and put AI into operation to make business processes even smoother and more customer-centric."

IT security and privacy

The AI Factory is based on Amazon Web Services (AWS). This means that basically any business unit that wants to use corresponding AI services can use cloud computing to its full extent and in a fully compliant manner. The high requirements and strict specifications in Europe form the clearly defined framework. For ERGO, this means: the highest level of IT security, consistent compliance with binding data protection guidelines with regard to personal data as well as multi-client capability and separation of different user roles are guaranteed at all times. ERGO has also set itself ethical guidelines for the use of AI, which are taken into account when implementing use cases.

This design, which is both flexible and technically and professionally compliant, makes docking processes to existing systems simple. Various tools and services specialised in AI offer data analysis teams and data specialists maximum effectiveness. In short: as a state-of-the-art IT platform, the AI Factory simplifies the implementation of various artificial intelligence models that will have a digital impact in many places in ERGO in the future.

Clear decision for cloud solution

Even though parts of the AI Factory are operated in the company's own data centres in order to implement interfaces to the ERGO core application systems, the core of the AI Factory is located in the cloud. The infrastructure based on this scales automatically, integrates the latest (data) technology and is equipped with components that allow continuous optimisation. This allows developers and data specialists to quickly and easily create and offer models for a wide range of applications without having to touch the infrastructure itself. Investments in expensive infrastructures, high licence costs in advance and ongoing maintenance fees are no longer necessary, as only the pay-per-use model is used. Another advantage of this method: overcapacities and superfluous computing resources are eliminated. This makes it possible to use new functionalities or simply more computing power on an ad-hoc basis without having to buy new hardware. This model is therefore significantly more sustainable than operating exclusively in ERGO's own data centre.

The cloud-based infrastructure allows components to be set up in a secure, efficient and controlled manner. Once developed, AI models do not have to be deployed from scratch each time. In addition, the use of cloud services facilitates the updating of AI models already in use as well as their ongoing control and monitoring. This is because prefabricated and configured solution modules are already available today and can be expanded and supplemented as needed. To prevent security threats and anomalies in the highly complex and dynamic processes, the AI Factory relies on cloud monitoring solutions. Every (data) movement is continuously assessed and measured.

Consolidation and increased reach

At the end of last year, the AI cloud platform was put into production for both model development and the operationalisation of use cases. We will report in more detail on exemplary AI use cases of the ERGO AI Factory in a second part. The prerequisites are currently being created to make the AI Factory available to ERGO's international subsidiaries. However, the possibilities are far from exhausted. The comprehensive and flexible approach of the AI Factory also promises to support and meet the needs of other industries for the use of artificial intelligence in business processes.

Further information on the AI Factory can be found at: www.ergo.com/aifactory

Text: Martin Sulkowsky