Artificial Intelligence

Superhuman Assistant

Predictive maintenance

Sebastian Fetz
18 Jul 2023
14 min read

The Vision of the 'Superhuman Assistant': Going Beyond Human Limits with AI – or Why Integrating Large Language Models is Essential for Corporate Strategy and Guidelines for Effective Execution

Names of people like Isaac Newton, who fundamentally changed the foundations of physics with their groundbreaking theories, shine in the annals of scientific history. Behind these extraordinary geniuses were often less well-known individuals with astonishing abilities, known as savants. An example is Rüdiger Gamm, a German savant and math prodigy who can solve complex calculations that would challenge even the most powerful computers in a matter of seconds. But what if every organisation had access to such "superhuman" capabilities?

Imagine an assistant who carries all of a company's information in her "brain". She recalls the balance sheet from 1998, the underlying data, the marketing strategy for 2030, and the contents of today's newspapers as well as all newspapers published since documentation began. She tracks everything posted on social media channels. And she is not just a repository of data - she can use all this information for complex analyses. The answer to this futuristic vision could lie in the world of Large Language Models (LLMs).

LLMs are artificial intelligence models capable of analysing, understanding, and generating massive amounts of text data. They can perform analyses in seconds that would normally take specialised teams days, weeks, or even months, thereby providing an obvious conclusion for saving time and money.

In an audacious forecast, Goldman Sachs suggests that generative AI has the potential to boost global GDP by an impressive 7% - a notable impact for a single technology. But these potentials are not merely the stuff of science fiction or visions of the future. With today's Large Language Models (LLMs), much of what sounds like fantasy is already possible. And the models of the future will be capable of conducting even more complex analyses.


"The Race Is On"

In a wave of technological pioneering, visionary market leaders such as Telefónica, Mercedes, and Bloomberg are already harnessing the immense power of Generative Artificial Intelligence and Large Language Models (LLMs). This early embarkation into a burgeoning era of innovation is driven by four key motives that, like invisible force fields, dictate the strategic direction.

Racing Against Time: Innovation Cycles in Fast Forward
In the pulsating world of digital technologies, the innovation cycle resembles a rapid time-lapse. Those who fail to keep up quickly fade into the fog of irrelevance. By boldly leaping to the forefront of technological development, companies secure decisive competitive advantages and a pole position in the digital transformation through the early implementation of Generative AI and LLMs.

Hunting for the Brightest Minds: The Battle for Talent
Like in an intense competition over the brightest minds of our time, the hunt for the best talent has begun. This competition is reflected in the rapidly increasing number of job advertisements for AI experts. Companies that recognise and utilise the potential of Generative AI and LLMs present themselves as attractive employers and attract the brightest stars in the talent sky.

Industry Upheavals: The Disruptive Power of Generative AI
Generative AI is a disruptor, a ground-shaker with the potential to turn entire industries upside down. Even tech giants like Google find themselves confronted with the disruptive power of this technology, as demonstrated impressively by the rising company OpenAI. In a chess game of the giants, the moves are being redefined and the balance of power is being recalibrated.

Customers in the Age of Automation: Changing Expectations
In a world where automation is becoming the norm, customer expectations are shifting. They expect not only lower prices but also an improved customer experience. Virtual assistants, chatbots, and similar technologies, powered by LLMs, are setting new standards in customer interaction. Companies that fail to meet these transformed expectations risk losing touch with their customers.

The starting shots for the race for supremacy in Generative AI have been fired. Who will ultimately lead the pack and how the technological terrain will continue to shape is yet to be seen. But one thing is certain: those who now bravely pedal and actively set the course for the future have the best chances of reaching the finish line as winners.


The Benefits of Generative AI and LLMs

1. Increased Efficiency and Cost Savings: Generative AI and LLMs can automate complex tasks, freeing up employees from time-consuming activities and allowing them to focus on more strategic and creative tasks. According to a McKinsey report, up to 50% of current work activities could be automated between 2030 and 2060. This could lead to significant personnel cost savings and simultaneously increase productivity and efficiency within the company. With this timeline rapidly approaching, this underlines the necessity to strategically implement Generative AI and LLMs now.

2. Improved Decision Making: These models can quickly and accurately analyse vast amounts of data, providing insights that support strategic decision-making. They can help CEOs and their teams make data-driven decisions, reducing uncertainty and risk.

3. Enhanced Customer Service: LLMs allow for personalised and efficient customer communication around the clock. This improves the customer experience, which can, in turn, increase customer loyalty and revenue.

4. Information Advantage: Through LLMs' analysis capabilities, companies can glean valuable insights from their data and make informed business decisions. They can also predict current market trends, thereby gaining a competitive advantage.

5. Driving Innovation: By utilising Generative AI and LLMs, companies can develop new business models and services based on personalised and interactive customer experiences.

6. Scalability: Generative AI and LLMs enable companies to scale their operations without a proportional increase in resources. As AI models can handle larger volumes of data and tasks as they scale, companies can expand their services or customer base while maintaining high efficiency and customer service.

7. Risk Reduction: AI models can be used to identify and mitigate risks in real-time. This can range from detecting fraudulent transactions in the finance sector, predicting device failures in manufacturing, to identifying potential PR issues in social media data. By proactively managing these risks, CEOs can protect the reputation and financial health of their company.


Strategic Considerations for Implementing Generative AI and LLMs

Many companies, even large ones, in our view, make the mistake of investing in Generative AI, but only at a very low level via master's students, working students, and interns. For companies wishing to fully harness the incredible power of LLMs, they should give these technologies strategic priority. Several important aspects should be considered:

Strategic Alignment: LLMs should not be viewed as just another technology, but as a central component of corporate strategy. The integration of LLMs should be closely linked to the company's goals and purpose and reflect the value that this technology can deliver.

Data Privacy and Compliance: The implementation of LLMs requires careful consideration of data privacy and compliance. It is essential that companies ensure the use of LLMs complies with legal regulations and best practices, especially when dealing with sensitive information.

Building Competence and Training: Companies should build internal expertise as well as draw on external expertise to get the most out of LLMs. This means employees not only need to be trained to effectively use the models, but also to understand how they can be applied to various business processes and challenges.

Flexibility and Scalability: The infrastructure and processes used to support the LLMs should be flexible and scalable. They should be designed to adapt to the changing needs of the company and advances in LLM technology.

Change Management and Cultural Shift: Implementing LLMs is not only a technological challenge but also a cultural one. Companies need to be ready to implement change management strategies and foster a culture of acceptance and adaptability to ensure the success of this revolutionary technology.

Ethics and Responsibility: Like any technology, LLMs can be used for good and bad. It is important to incorporate ethical considerations from the start of the implementation process and ensure responsible use of LLMs. For instance, it might be important to develop ethical guidelines for the use of LLMs and ensure that these models are not used for manipulation or the dissemination of misinformation.

User Experience and Acceptance: For the successful deployment of LLMs, it's crucial that end users - whether company employees or its customers - accept and use the technology. Therefore, the focus should be on creating a positive user experience, such as through intuitive design and user-friendliness of applications based on LLMs.

Evaluation and Continuous Improvement: As with any technology, it's also important with LLMs to continuously evaluate their performance and benefits. This includes setting appropriate metrics for measuring success and establishing feedback loops to continually improve the models and their application.

Interdisciplinary Collaboration: When implementing LLMs, it can be advantageous to involve experts from various fields, including data science, linguistics, ethics, law, and business administration. Such interdisciplinary collaboration can help better understand and address the challenges and opportunities associated with LLMs.


Strategic Planning and Levels of Maturity

Perelyn, we position our customers, in collaboration with them, within a 5x5 matrix that correlates the five phases of IT/AI infrastructure maturity (Initial, Managed, Defined, Quantitatively Managed, Optimising) with the five stages of Generative AI strategy maturity (Awareness, Understanding, Capability, Proficiency, Leadership). This matrix could serve as a roadmap that showcases the company's progress in terms of technical infrastructure and the ability to implement and utilise generative AI models. It can help determine the current state of the company and plan the next steps for improving the usage of LLMs. With such a structured approach, it can be ensured that the company continuously evolves and fully leverages the enormous opportunities that LLMs offer.

Stages of IT/AI Infrastructure Maturity

1. Ad-hoc: The organisation has little to no formalised IT/AI infrastructure. IT operations and services are improvised or reactive. There are no formal procedures for handling IT issues, and there is little strategic thinking applied to the role of IT/AI in the organisation.

2. Emerging: At this stage, the organisation starts to recognise the significance of IT/AI and integrates it into its operations. Basic infrastructure may be present, but it's not comprehensive or well-managed. AI understanding is limited, and use-cases are isolated.

3. Defined: The organisation now has a proper IT/AI infrastructure with formalised procedures and responsibilities. IT/AI is viewed as a critical component of the business strategy. There's increased focus on data management, security, and compliance. Regular training and updates on AI capabilities take place.

4. Managed: The organisation possesses a fully developed IT/AI infrastructure that is regularly updated and improved. The infrastructure is integrated into all areas of the organisation, and there's an understanding of the potential impact of AI across the organisation. Decisions are often data-driven, and concerted efforts are made to improve AI skills and knowledge.

5. Optimierung: At this stage, the organisation continuously improves and optimises its IT/AI infrastructure. The organisation leverages AI to drive innovation and competitive advantage. There are robust standards for data management, privacy, and ethics. The impacts of AI are regularly evaluated, and there's a profound organisational understanding of AI.


Stages of Generative AI Strategy Maturity:

1. Exploration: At this phase, the organisation is just beginning to understand the potential of Generative AI. Simple models may be tried out, or use-cases identified where generative techniques could make a difference.

2. Proof of Concept: The organisation has identified specific problems that can be solved with Generative AI and develops prototypes or proof-of-concept models. There's an understanding of the potential value, but the application or integration is still limited.

3. Operationalisation: Generative AI models are implemented in production environments. The organisation starts to see tangible benefits from their models, but there may still be challenges in scalability and adaptability to new data and requirements.

4. Integration: Generative AI is seamlessly integrated into the organisation's processes and workflows. The models not only function but improve over time, learning from new data and providing continuous added value.

5. Innovation: In the final stage of maturity, the organisation leverages Generative AI not only to enhance existing processes but also to innovate and create new opportunities. Generative AI drives strategic decisions and is seen as an integral part of the organisation's competitive advantage.

The applicability of the cells in the 5x5 matrix to companies of different sizes depends heavily on the specific circumstances of each company, including its current technological maturity and strategic orientation. However, here's a general assessment:

Small Businesses: Small businesses are likely in the early stages of the matrix. They may have a basic IT/AI infrastructure (Stages 1–2: Initial, Managed) and are either aware of the potentials of generative AI models or are just beginning to understand them (Stages 1–2: Awareness, Understanding). Hence, the cells representing this intersection would be best suited for small businesses.

Medium-sized Businesses: Medium-sized businesses might have a more advanced IT/AI infrastructure and be capable of leveraging certain AI functionalities (Stages 2–3: Managed, Defined). They might have a better understanding of generative AI models and are building their capabilities in this area (Stages 2–3: Understanding, Capability). Therefore, the cells representing this intersection would suit medium-sized businesses.

Large Corporations: Large corporations, especially those in technology-intensive industries, could be at the higher stages of the matrix. They might have a well-defined and quantitatively-managed IT/AI infrastructure (Stages 3-5: Defined, Quantitatively Managed, Optimising) and are on the path to mastering or leading in generative AI strategies (Stages 3-5: Capability, Proficiency, Leadership). Thus, the cells representing this intersection would suit large corporations.

The journey towards implementing Large Language Models (LLMs) is akin to venturing into a new era, an exploration of uncharted territory that's both rich in possibilities and laden with challenges. If we envision savants like Rüdiger Gamm, who can calculate with impressive accuracy and speed, we see a picture of the enormous potential that can be unlocked through the fusion of human intelligence and technological innovation.

The "Superhuman Assistant," this nearly mythical notion of artificial intelligence capable of learning, adapting, and collaborating closely with us, suddenly doesn't seem quite so utopian anymore. Every stride we make in AI research brings us closer to this goal.

The introduction of LLMs may seem like stepping into a new era. Yet, as the work of savants like Rüdiger Gamm demonstrates, the amalgamation of human potential and technological innovation can yield groundbreaking results. The "Superhuman Assistant" might soon cease to be a fantasy, transforming into a valuable resource for businesses across the globe.


Discover Generative AI with Perelyn: Your Full-Service AI Consultancy

The world of generative Artificial Intelligence is vast and uncharted - but we at Perelyn are your experienced guides on this thrilling journey. Begin your expedition with our complimentary Starter Phase, structured in three clear, grounded steps:

1. Broad Overview: We offer you a comprehensive overview of the current situation, developments, and innovation trends in your industry and amongst your competitors.

2. Readiness Check-up: We evaluate your current status regarding IT/AI infrastructure and the readiness to deploy generative AI.

3. Personalised Route Mapping: We deliver a customised roadmap for the next 3 to 12 months, aligned with your specific needs and goals.


Why Perelyn is the right partner for you:

Extensive Project Experience: We have completed over 20 successful GenAI projects across diverse industries (e.g., Logistics, Media, Manufacturing, Pharma) and functions (Marketing, Customer Care, R&D), and overall, more than 200 AI projects for medium to large companies.

Cost-Efficient with Near-Shore Solutions: Our adaptive pricing model and the ability to leverage near-shore capabilities help you efficiently manage costs.

Pool of Experts: We have a network of over 50 experts in the DACH region, Scandinavia, and Eastern Europe, ready to steer your project to success.

Strategic Pioneer in Generative AI: We are the first company in Germany offering a comprehensive strategic approach to generative AI.

Thought Leadership and Industry Support: As a leading thinker in the field of generative AI and Large Language Models, we are recognized and supported by leading companies and organizations.


Take the first step into the future with Perelyn - we look forward to the joint exploration journey into the world of Generative AI.

Other Insights

How do we solve problems?

Please consider helping us improve our website. By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation and analyse site usage. View our Privacy Settings for more information.