Capgemini: Rise of agentic – AI How trust is the key to human-AI collaboration

Capgemini: Rise of agentic – AI How trust is the key to human-AI collaboration

The AI Reputation Monitor 2025 by FTI Consulting shows how strongly large language models such as ChatGPT are shaping the perception of companies and managers. In the analysis of the DAX40 companies, 83% exhibit at least one reputational risk in the AI-generated responses. These are often based on older or outdated information that remains present in the AI logic or is interpreted incorrectly. The simplified representation of complex topics by the models in particular can reinforce problematic narratives.

The influence of CEOs on the AI-based image of their companies is particularly striking. In many cases, their profiles have a direct impact on the overall perception – both positive and negative. Visibility, tonality and relevance in digital channels play a central role here.

The study highlights the type of content that is particularly influential in AI responses. In addition to traditional media reports, this primarily includes well-structured, machine-readable information from corporate websites and profiles on platforms such as LinkedIn. Content with context, classification and concrete utility value has a much stronger influence on the statements generated than mere headlines or advertising messages.

Several strategic options can be derived from the results in order to shape the digital reputation profile. These include transparent counterstatements, the creation of consistent content in AI-relevant formats and the regular review of typical prompt responses.