In the ever-evolving landscape of artificial intelligence, we find ourselves at the crossroads of innovation, where the convergence of rule-based AI and generative AI is reshaping the way businesses operate. In the world before OpenAI’s ChatGPT moment, the general understanding of AI was that of a driver for automation. But more and more new generative AI applications and augmentation applications are taking center stage. While both types of AI on their own can have a massive impact on society and the economy, the really big impact comes from a combination of the two.

/ AUTOMATION – RULE-BASED AI DRIVES EFFICIENCY GAINS

Rule-based AI, with its deterministic nature, has long been a cornerstone in solving specific problems with predefined rules. Its ability to follow a set of human-coded rules and guidelines with precision and predictability has made it indispensable in industries where compliance, safety, and well-defined processes are paramount. The expected outcomes of these processes are predefined, automating routine tasks, or ensure regulatory compliance in suitable use cases. In this area, use cases can often be implemented with ‘off-the-shelf’ AI products.

Perhaps the most well-known example of the underlying technology is Robotic Process Automation (RPA), which is widely established in Finance. In this context, RPA can download PDFs from a database, convert the document into a new format, extract predefined fields, and add the information to a new file that can be input for other processes. Costs savings of up to 80%, compared to a human performing the same job, are not unusual in these cases. RPA operates on a simple yet effective ’cause-and-effect’ methodology and requires only basic data and information to operate successfully. The technology has been widely adopted and is not as novel as generative AI.

/ AUGMENTATION – GENERATIVE AI UNLEASHES CREATIVITY AND ADAPTABILITY

On the other hand, Generative AI redefines the collaboration between humans and AI. The creation of (high value) text, images or videos have predominantly been associated with humans. The technology, it should be noted, is not brand-new. However, it has made enormous progress in the last year and will continue to do so. In this area, use cases are typically implemented via project approaches, but a trend towards increasing productization can be observed.

Machine learning enabled by the concept of transformer made it possible to train the AI with ever-larger models without the need to label all of the data in advance. Based on the available training data new content can be created. Specific AI providers are still numerous; however, we already see the beginning of a consolidation in the market with players like OpenAI soon launching their GPT marketplace and Google’s new Gemini Ultra AI model. Thousands of intangible examples of generative AI can easily be found on LinkedIn and Co. An existing and rather tangible example is in the Nike House of Innovation, where customers can design their own shoe or test their new clothes in a cinematic product try-on. In contrast to RPA, generative AI works with large language models, also known as foundational models, as they are the building blocks for various natural language processing tasks. They require a large amount of training data as input in order to improve their capabilities and generate content that is equivalent to or better than human-generated content.

/ THE POWER OF SYNERGY – RULE-BASED MEETS GENERATIVE

The true magic happens when we bring these two paradigms together. Imagine a scenario where the precision of rule-based AI is complemented by the creativity and adaptability of generative AI. This combination is a game-changer, regardless which AI finally seems to be customer facing. Both types of AI have their weaknesses and strengths and the combination of both balances out the weaknesses and makes their strength even stronger. For example, generative AI’s ability to understand large sets of language can help clean up the data that rule-based AI requires to operate or is supposed to process.

Mercado Libre, the Amazon of South America, uses AI to increase the overall efficiency of the AI applied in the company and has implemented GitHub Copilot, an Artificial Intelligence that writes software code and fixes identified bugs intelligently (Augmentation). The Copilot can even write a Chatbot which helps Mercado Libre to accelerate their handling of customer enquiries. At the same time RPAs (Automation) are struggling with a limited performance due to poor data quality. Here, Generative AI can support the data cleansing and data correction ambitions of Mercado Libre, eventually increasing performance and customer satisfaction.

In our customer projects, we come across a wide range of AI use cases – from simple, ‘low-hanging fruit’ to highly sophisticated and complex use cases. In our experience, some use cases, which ‘intelligently’ combine automation and augmentation, have quickly had a major impact on businesses and people:

• Customer Support: Rule-based AI can handle routine queries with efficiency, while generative AI can understand and respond to nuanced and unique customer inquiries, delivering a personalized and human-like interaction. O2, a brand of Telefonica, has created Aura which is a chatbot that can support customers with special inquiries, such as providing information about the contract or providing information of service disruption.
• Compliance and Risk Management: Rule-based AI ensures adherence to regulations and predefined policies, while generative AI analyzes and generates insights from unstructured data, helping identify emerging risks and opportunities. The company Sirion offers a smart contracting platform, which creates new contracts based on generative AI but proactively tracks renewals and expirations based on predefined rules.

• Content Generation: Combining rule-based guidelines with generative creativity results in dynamic content generation, from marketing materials to product descriptions, tailored to the audience and context. Anima provides a platform for developers to seamlessly translate design into code. At its core is a rule-based code generation engine based on predefined algorithms and heuristics that provides a stable, efficient and predictable code generation process. The company’s breakthrough innovation is the integration of LLMs based on the non-generative AI code. This combination allows the system to improve and iterate on the base code, resulting in more efficient and creative code output that is well suited to the clients’ needs.

Concluding, we want to emphasize that while experimentation with separate use cases in automation or augmentation is crucial, we suggest to already include AI use cases in your AI project portfolio that combine augmentation and automation.

Due to the age of rule-based AI, learnings are already vastly available. In order to bring it to the next level, it should be combined with generative AI to harness the strengths of both and create a more versatile artificial intelligence system.

To fully harness the potential of AI, organizations should embrace the synergy between rule-based and generative AI, integrating both into their project portfolios. By combining the efficiency of rule-based automation with the adaptability of generative AI, businesses can unlock transformative opportunities, driving innovation, enhancing customer experience, and improving operational performance.

The time to elevate AI initiatives and leverage this powerful combination to gain a competitive edge is now.

/ ABOUT THE AUTHOR

• Dr. Carsten Linz is the CEO and Founder of bluegain. Formerly Group Digital Officer at BASF and Business Development Officer at SAP, he is known for building €100 million businesses and leading large-scale transformations affecting 60,000+ employees. He is represented on various boards including Shareability’s Technology & Innovation Committee and Social Impact. A member of the World Economic Forum’s Expert Network, Dr. Linz is also author of renowned books and articles who shares his expertise in executive programs at top business schools around the world.