Workshop on multi-agent systems with EY and Leaders of AI

A look behind the scenes of our multi-agent system workshop

Artificial intelligence is changing the world of work rapidly. But how quickly can companies really benefit from AI? Our joint project with Leaders of AI, EY and other partners provides surprising answers.

The starting point: new technology, many open questions

Multi-agent systems, that is AI teams that work together independently, are still new territory for most companies. The central questions we asked ourselves:

• Can AI agents really function as a team?

• How can human ways of thinking and decision logic be transferred to AI systems?

• Which practical use cases deliver real business value?

We did not want to discuss these challenges in theory but to solve them in practice.

Strong partners for an ambitious project

For this project, five companies with complementary strengths joined forces:

Leaders of AI brings expertise in AI training and prototyping formats and enables specialists and managers to use AI effectively.

EY (Ernst & Young AG) as a global consulting firm contributes a deep understanding of business model transformation and innovation in regulated markets.

ThynkAI develops scalable AI solutions that automate knowledge, processes and decisions, with a focus on MLOps best practices and adaptive applications.

Virtual Identity AG took on the strategic workshop design and orchestrated the complex event with a wide range of stakeholders.

scoreprise.AI (us) contributed our technical expertise in building multi-agent systems and supported the participants with the practical implementation.

Months of preparation meet one day of innovation

The key to success lay in the preparation: over several months, we jointly developed a functional prototype of a multi-agent system. This was to serve as a solid basis for the workshop.

The scenario: a fictional company needs an AI team for market monitoring in the social media and marketing area. The hierarchical structure comprised an AI CEO who makes decisions and delegates tasks to specialised AI employees, similar to a real company.

On 21 February 2025 the time had come: at the EY Wave Space, 60 experts from a wide range of areas came together, prompt specialists, management professionals, AI developers and marketing experts. Divided into five groups, they were given the task of developing the base prototype further and adding individual specialisations.

The decisive breakthrough

What surprised us the most: the speed of development exceeded all expectations.

While many companies plan AI projects as lengthy processes over months or years, our workshop demonstrated the opposite. In just one day, five different, specialised multi-agent systems were created, each with its own focus and its own character.

The most important insight: what is needed is not primarily deep technical expertise but a heterogeneous mix of human skills from different disciplines. The combination of a prepared technical foundation and interdisciplinary collaboration made the difference.

Concrete results that convince

At the end of the day, all five groups presented their individually developed multi-agent systems for market monitoring and content strategy development. The best solutions were awarded prizes and impressively showed the variety of possible application scenarios.

The key insights:

• AI agents can already work together as a team today and coordinate with each other

• With the right preparation, functional AI teams can be developed in a very short time

• Well-known companies such as Red Bull showed concrete interest in implementing such AI departments

• The hypothesis was validated: in future, every company will use AI employees, not only in collaboration with people but also with each other

What does this mean for your company?

Our workshop showed: the technology is ready. Multi-agent systems are no longer a distant vision of the future but can be implemented quickly with the right expertise and preparation.

The decisive success factors are:

1. A solid technical foundation through experienced partners

2. Interdisciplinary teams with different perspectives

3. Practical use cases with measurable business value

4. An iterative approach, from prototype to specialisation

The successful collaboration between Leaders of AI, EY, Virtual Identity, ThynkAI and scoreprise.AI demonstrates how, by combining different areas of expertise, innovative AI solutions can be developed in record time.

Interested in multi-agent systems for your company? Contact us to find out how we can jointly develop AI teams that create real added value.