Potentials and limits of AI: Current and future use cases

Generative LLMs such as ChatGPT, Bard or LLaMA use advanced deep learning techniques and huge data sets to recognise, interpret and generate human language and complex data. The impressive results in text generation and other creative areas are not only milestones for the optimisation and digitalisation of existing processes, but also make a decisive contribution to the transformation of new areas of application by automating them and thus fundamentally changing the way we work and communicate.

Many companies are currently still in an experimental or pilot phase, trialling the latest AI applications in various business areas to test their efficiency and effectiveness. Nevertheless, their transformative power will bring about profound changes in the way companies work and communicate in the long term.

Five examples of current and future AI use cases

Below are some examples of how AI can be used in various industries and business areas today and in the future to optimise operational processes and enable new business models. These examples range from general applications to specific niche applications that demonstrate the versatility and potential of AI.

1. Analysing and evaluating large amounts of data

The use of AI for analysing and processing large volumes of data opens up a wide range of possibilities and applications that go far beyond the limits of traditional methods: For example, it can help to better design individual services and products. For example, the AI sub-area of natural language processing (NLP) can already be used in the insurance industry to analyse patterns and trends in data and compare documents at content level. Changes to general insurance conditions can be recognised in this way or an insurer’s services can be compared with the current market offer. Due to their complexity, these tasks are particularly time-consuming and error-prone when performed manually. With machine support, complex documents can now be automated, analysed and compared at a semantic level, significantly reducing errors and effort.

2. Customer communication in specific industries

One example of a specific niche application is the use of AI in customer communication. The use of large language models in particular offers an improvement in customer interaction through natural language dialogue systems. In sectors such as banking, insurance or telecommunications, this can help to process customer enquiries more efficiently in future through a high degree of automation and offer hyper-personalised advice and customer contact. Well-trained LLMs can respond even more individually to customer needs, suggest suitable products or services and feed findings back to the company in order to collate this customer information in compliance with data protection regulations.

3. Enterprise asset management and predictive maintenance

A special use of AI can be found in the area of enterprise asset management. AI is already being used in industry today to predict and carry out maintenance before faults occur. This predictive maintenance is based on the analysis of sensor data and other relevant information in order to recognise patterns that indicate possible equipment failures. In this way, specialists are informed before a fault occurs that a particular component is worn and should be replaced. Intelligent image processing systems are another field of application for AI in industry. These support quality control by automatically recognising errors in certain manufacturing processes.

4. Software development

In the field of software development, the use of AI will help to speed up development and make it more efficient in the future. Application scenarios include the automation of routine tasks, the generation of standard code or the use of large language models (LLMs) as intelligent tools for answering technically demanding questions.

5. Internal use to build a knowledge database

A niche application is the creation of internal knowledge databases. Already today, chatbots such as ChatGPT can be enriched with industry-specific and proprietary knowledge as well as data held within the organisation to create an internal knowledge base. This supports specialists in answering specialised questions.

Current challenges and limits of AI

Investing in innovative AI-based systems can give companies a significant competitive advantage by using data-driven insights and automation to optimise processes or open up entirely new areas of business. However, these advances also bring with them ethical and data protection challenges as well as security risks that need to be addressed appropriately in each use case.

AI applications are only ever as good as the quality of the models and training data they use. Large language models in particular are currently dependent on a huge amount of high-quality training data, which could soon run out. At the same time, these new generative AI applications still raise unresolved questions regarding data protection and privacy rights as well as compliance with the General Data Protection Regulation (GDPR), which are currently being examined by the European data protection supervisory authorities.

AI hallucinations are also a known problem of generative language models, where the chatbot provides incoherent or incorrect answers. Risks also lie in possible distortions with regard to criteria such as origin, gender or other characteristics that can be reproduced based on historical data and the prejudices contained therein.

In the field of software development, there are currently concerns about AI-generated software being used incorrectly without human supervision and expertise. This is because poorly written code by AI systems could cause more damage than help solve problems in the future due to the resulting potential security vulnerabilities.

Human expertise remains essential

The current challenges and technical limitations make it clear that artificial intelligence is a valuable tool to support specialists in companies in their work. However, in critical or sensitive use cases, such as the evaluation of promotions, accounting and controlling processes or claims settlement in the insurance sector, which have a significant impact on customers, decisions should not be left exclusively to AI. For business-critical decisions, human expertise and experience remain essential to safely and effectively utilise the current capabilities of AI. This connection between human and machine is crucial for business transformation as it increases efficiency while ensuring the quality of decisions.