Times are turbulent and will be staying this way. Technologies represent the strongest drivers of increasing speed and the dynamics of change. After internet, smartphone and the resulting platforms that have been created, the topics around artificial intelligence (AI) are currently the biggest trigger and at the same time a source of hope for the economy and society. AI has the potential to cover almost all areas and activities of our lives through automation of work steps, processing of large amounts of data, certain predictions and making decision-making processes more efficient. Despite the enormous potential, as entrepreneurs, networkers and investors we always see three challenges in the application of AI:
1. AI is a lot of “theater” - Often there is less AI in it than expected
With new promising technological leaps, a lot is often said about the technology itself but too little about the real meaning and even less about concrete applications. The keyword AI can be found in many pitch decks, on websites etc., without really applying AI. MMC Ventures recently uncovered that in 40% of all AI start-ups in Europe there really is no AI in them. This is a clear indication of a lot of theater with little substance. We see the intelligent machine processing and visualization of data based on statistic mathematical methods as one of the core services of modern analytical approaches like AI. For this, enough data points are needed. These must be ordered, sorted and made comparable.
2. Understanding adaptation gaps
To understand the current technology development, “Moore’s Law” is often used as an explanatory construct. According to this principle, AI as a technology develops exponentially, i.e. the technological performance is doubled regularly. However, the problem is that there is a gap between what is technologically feasible and the reception or adaptation by the individual. The gap, also called adaptation gap is getting bigger when it comes to the adoption of technology by (business) organizations and is even bigger regarding public institutions. AI as a technology offers many possibilities - but the people and organizations are at the same time not able to use them. Still, the end user can adopt it relatively quickly. AI is integrated in many applications in daily life and often runs in the background. However, companies continue facing major challenges regarding necessary structures and competencies.
3. Too many solutions for too few problems
One of the basic problems that we face repeatedly, despite all the technological euphoria, that there are simply too few specific questions or problems that are worth to be solved with AI. Identifying these problems is crucial. They are often less obvious because they are hidden in processes or niches and additionally are still very complex. For an outside (AI) developer, it is therefore very hard to get a proper insight. It always takes a suitable problem owner from the industry who deals with the problems, procedures and so on. Based on these problems we need to then look at what technological solutions can really help. Here we often experience that a simple business intelligence (BI) solution is usually already sufficient to solve the problem and at the same time create a business benefit.
In summary, there is a lack of concrete business problems that are easy to solve with advanced technology and at the same time are comprehensible by the end user. In this context deep tech as a category of AI comes into play. Although the term is also broadly defined, it is used far less inflationary compared to AI. In addition, deep tech stands for the ability, with the help of the latest scientific findings, to solve concrete business problems and thus to create innovations in the form of start-ups, amongst other things.
To further illustrate the possibilities of deep tech, we use the example of the Hasso Plattner Institute (HPI), which is seen as the talent foundry for software developers in Germany. Over the last few years, a cluster of researchers at HPI, led by Prof. Dr. Jürgen Döllner, has developed around the topic of visualization of data in real time. With a team of 30 employees he works on fundamental technological software solutions around the topic of deep learning and data visualization. The technologically feasible is matched with industrial partners and experts with concrete problems. From this, the company builder German Deep Tech has created numerous start-ups over the last years which have been successfully operating in the markets.
Two of them are introduced briefly to give an insight into what really is viable and which role deep tech in combination with AI can play to solve concrete problems.
Technology: AI-based location intelligence platform Customers/users: established industry and retailers with locations/shops, public transport companies, real estate businesses
Addressed Problems: Location-related data and transaction data have a big influence on the success of companies and organizations. To date, it was very laborious to involve such data in decision-making process or only possible through specialists and complex tools. As a result, a lot of data crucial to success in the decision making are not considered - there are no reliable “data-driven” decisions possible.
For real estate properties, specific analyses can cost so much time that as a result the property is not available anymore. In the retail sector the factors influencing success or failure of business locations are usually only insufficiently known.
Solution: Targomo offers a location intelligence platform that enables users to carry out location analyses without the need for special GIS (geographic information systems) knowledge. The location intelligence platform analyses geo-related data using artificial intelligence in real-time and uncovers complex dependencies. This way users receive valuable insights to increase sales, reduce costs and to achieve decisive competitive advantages.
The platform is already being successfully deployed domestically and abroad: Real estate and retail companies forecast the potential of locations to optimize branch networks and delivery routes; public transport companies adjust routing according to demand.
Technology: AI-based analytics platform for managing software development organizations Customers/users: Companies from all industries, whose competitive advantage is based on self-developed software (banks, insurance companies, mechanical engineering, trade, logistics, …)
Addressed problems: Software forms the basis for the business success of almost all industries – particularly, in the age of digitalization. However, managing internal software development departments or corresponding third-party service providers in such a way that software releases are efficient, of high quality and completed within the planned timescale, and made available to users, is a great challenge. In particular, regarding the subject of efficiency optimization, the software engineering discipline is still decades behind other engineering disciplines such as mechanical engineering. It is not uncommon for productivity losses to be so high in the development process that from 100% of invested development budget only about 20% in new features (business value) can be returned to the users. A reason for those elevated time and money losses are, for example, software errors and their clean-up in code. The subject is currently essential for business decision-makers as they are aware of the chances of digitalization in order to be able to lead their company to long-term success.
Solution: Seerene offers an end-to-end analytics platform that analyses the entire software process. This includes internal software development as well as those outsourced to third parties. Persons responsible can, for example, transparently identify in which areas the developers have to spend a lot of time on defect fixing or where there is “unclean code”, which may compromise current releases. Seerene makes available analyses, that are visually and easily understandable, in so-called software maps and in the form of a Digital Boardroom. Using this, corporate managers can minimize risks in software development, improve the quality and raise productivity by 30 percent and more.
In conclusion, we are convinced that technology will always prevail. The only open question is when and how. AI and algorithms already influence many areas in our lives today. In our view, deep tech is a concretization of the application of AI. Each (new) technology needs a new problem. The path to the solution can be differently designed - there is no patent remedy. But there are always great opportunities, particularly when problem owners (industry) and solutions (start-ups) connect and merge.