Fully digital companies such as Google or Amazon are also changing markets. They are not active because they shift the expectations that customers have of a company across all industries: Customer service is available around the clock without waiting. Simple ordering and billing processes. Real-time information about the delivery status. A purchase and contract history in which all data is visible. Personalized address instead of faceless mass communication.
These are just a few aspects of these changed expectations. Those responsible in companies must react to this. It doesn’t matter whether one of the digital heavyweights is a direct competitor or not. Because disappointing customer expectations are a gateway for competitors. Competitors who take advantage of this
Digital companies build and build their entire business models on data. The consequence: They understand their customers better, react faster, communicate more convincingly. The handling of data creates competitive advantages that turn entire industries off their hinges. Companies that were born and grew up in this digital world have this data fixation in their blood. The optimal data flow determines the structure of the organizations and the design of the internal processes. It is more difficult for companies whose organization and techniques come from other times and follow different paradigms. Who have yet to become aware of the importance of data for success.
A learning process that is necessary for the long-term survival of many companies. Those responsible in companies should now face these issues because artificial intelligence (AI) is currently reshuffling the cards in markets such as the insurance industry. And the essential basis of AI applications is data.
Success Depends On The Data
AI processes are not new, but their use in companies is only just gaining momentum. They are tools that open up new possibilities for improving offers, services, and communication. Some application scenarios are widespread: AI technologies work in the background of speech recognition applications. The same applies to solutions that extract content from documents and prepare them for automatic further processing. A use case plays a role, especially in the insurance sector – for example, around applications or claims processing.
Language, writing, image recognition: these are typical applications in which AI shows its strengths. Appropriate processes and structures around data are the prerequisites for the successful use of AI. A data strategy is required.
A company’s data strategy defines the basic framework for handling data. It addresses topics such as the different types of data, their origin, the kind of administration, use, authorizations for access and processing, and deletion. The cornerstones of a strategy can be determined based on the so-called life cycle of data:
- Phase 0: The creation and generation of data
- The number of potential data sources cannot be overlooked. Whether business processes, IT systems, knowledge workers, customers, or machines: they all contribute to the constant data flow. Decision-makers have to carefully check which sources make sense to include them in the data processing process – under the vital stipulation that any information could be usable later.
- Phase 1: Managing data, recording and storing it
- Companies have different options: Either centralized storage, for example, in data lakes, or corresponding central storage systems such as document management systems or business applications. Or decentralized storage, for instance, in different designs, devices, and local devices. Both variants have other advantages, which will be discussed in more detail later.
- Phase 2: Sharing data
- From the place where it is created, data must go to the site where it is processed. Data flows between IT systems in a defined and structured form. Employees have access to portals, applications, or communication channels such as emails or collaboration platforms.
- Phase 3: Using and processing data
- Data often only becomes valuable and meaningful when it is correctly interpreted. This is done in different ways. For example, in creative processes, data analysis by experts, dedicated enrichment of, for example, market data, or in defined business processes, for instance, in the form of reports. Another possibility, which has become more and more critical in recent years, is the analysis of large and unstructured data and the resulting forecasts using AI methods such as machine learning.
- Phase 4: The deletion of data
- Even if the cost of storage space fell into the abyss, the deletion of data is an integral part of the life cycle. This can have regulatory requirements, for example, data protection reasons. Targeted cleaning also helps to improve data quality.
From a high altitude, these are the processes that companies have to observe when handling their data. The challenge lies in shaping the details: What skills does a company need to manage data effectively? Which functions are necessary? What does a blueprint for organization, technology, and processes look like? The aim is for companies to collect and process data according to uniform standards and, in particular, to use it for business purposes. In an industrial company, parts and raw materials flow through the production process, and at the end, there is the finished workpiece. In the future, companies and insurance companies will be built around the flow of data similarly. The planning center is the development of a structured process, for example, a data platform.
The term data platform is a combination of technologies, processes, and functionalities. It aims to enable the use of data in the company. In the forum, companies describe the structuring and networking of data-based processes and technologies: from the data sources at the beginning to implementing new AI-driven services or offers.
The creation of uniform structures in handling data addresses one of the central problems of data, or AI uses the juxtaposition of isolated silos. Contract documents are in the CRM system, and inquiries to customer service are in a separate application. Sales don’t know what marketing is doing – and vice versa. Automated personalization of the customer approach, for example, is hardly possible on this basis. The aim is to break down these silos, establish consistent responsibilities, and create incentive systems based on the quality of the data, for example. Only then do AI applications show their strengths.
How do companies manage to move from the current situation to a structure based on data requirements? Two ways lead to the goal – with different advantages and disadvantages.
On The Way To The Data
Approach 1: The parties involved implement processes that have been prepared for a long time and worked out in detail. You try to consider all eventualities in advance. After numerous rounds of coordination and years of planning, the data platform is ready for use.
Approach 2: The first AI use case is followed by the second and third. Those involved tap into data sources again for each project, design new processes, and use different data formats.
The descriptions of the two approaches are exaggerated, but the characteristics can be recognized in practice. The first carries the risk of over-engineering. There is a risk that the experts will quietly develop an ivory tower concept.
A platform that is prepared for every conceivable case. But it takes years before it works. And which then creaks at all corners and ends when used in practice. Because those involved can never foresee all the exceptions and unique features, on the other hand, the second approach focuses too much on the operational level without taking sufficient account of the strategic objectives of a data-driven company.
The quick implementation and the resulting immediate success are bought with a patchwork of individual measures. This approach works for a few use cases, but not on a large scale. Companies do not realize the advantage of “industrial” data processes with this approach.
When building a data platform, the middle ground between the two extremes is the right one. Keeping the big picture in mind while the individual project is being implemented is the right approach. That means: The construction of a data platform is a continuous, permanent, agile, and iterative process.
To achieve operational results and at the same time create strategic foundations, those involved have to coordinate a whole series of individual fields of action parallel to one another or at least closely timed one after the other. This includes the topics of architecture, processes, consulting/use cases, competencies, and organization. Especially in the first development projects, when the experience values are low, and the hurdles are high, the project team is confronted with numerous questions. But every new project ensures that new processes and working methods are embedded, that technologies and interfaces are available, and that the necessary skills are on board. Step by step, a data platform is created in which data flows largely automatically.
The aim is to use data profitably – and that means that a company has to adapt to the optimal flow of data in terms of its culture, its understanding of the value of the data, its organization, and its processes. It’s a time-consuming process, but it’s worth it – and in more and more industries, it is decisive for a company’s future viability.