HELDENKOMBINAT
In the quick hunt for productivity gains using AI, our “non-clients” usually realize only a fraction of the possible cost savings over the entire life cycle of the respective AI applications.
Our mission is to change this!
THE THREE MOST COSTLY MISTAKES WHEN USING AI
1
Poor conceptual understanding
Data preparation, tool selection and model training are complex, but useless if it turns out that the task cannot be solved (cost-effectively) by AI. For more than 80% of AI project ideas, it can be answered conceptually - i.e. without a technical proof of concept - whether the task at hand can be solved technically and commercially in a meaningful way by an AI application despite the very special requirements of IT security and data privacy.
2
No focus on reducing operating costs
Organizations have a strategic competitive advantage when achieving consistent quality at lower costs. The higher the operating costs of an AI application, the higher the potential to realize significant savings in the future through learning effects or technical progress. However, many AI vendors do not pass on such cost savings and many organizations are not positioned to easily replace AI vendors in a new tender.
3
“Error rates” in operation that are too high
Humans make mistakes, and so do AI models. The “error rate” of AI models leaves automation potential unused or, in extreme cases, can lead to users not accepting the entire AI application. However, the levers and effects of reducing such error rates, are well understood and can be estimated easily before the continuous improvement process starts.
THIS IS HOW WE SUPPORT OUR CLIENTS
Potential Analysis
Many AI projects arise by chance because innovation budgets or data are available or because AI providers (e.g. of generative AI) create huge sales pressure. We support our clients in systematically identifying, evaluating, and prioritizing AI potential to increase productivity in organizations and product applications.
​
​
Vendor Selection
The knowledge necessary for appropriately selecting the AI providers is not sufficiently available neither in corporate organizations nor in most consultancies and agencies. We support our clients in identifying suitable AI providers and understanding which AI technology and data are used by which provider and what implications this has for evaluating and selecting the right provider.
Implementation
Process digitalization and the use of new AI tools create a need for change at different levels and functional areas. We support our clients in identifying necessary changes in organizations and in coordinating and working through a realistic implementation plan together with all stakeholders.
Strategy Development
Many AI initiatives arise from an individual task or as a silo idea in organizations. We support our clients in systematically establishing and expanding economies of scale and competitive advantages through the cross-country or cross-functional use of AI and minimizing lock-in effects in the future.
Security Audits
For many organizations, AI is a new – and at the same time self-learning – toolbox with a black box character. We support our clients in understanding data bases and model architectures as well as identifying, evaluating and, if necessary, mitigating the resulting risks.
Chief AI Officer Coaching
The implementation of AI projects in organizations must be prioritized, promoted, and managed by a dedicated role or unit. We support our clients in growing into this role. We are a competent discussion partner and forward-looking advisor to cope with the rapid technical AI progress.
SOME OF OUR REFERENCES
Since 2016, we have evaluated more than 350 AI providers and their applications for corporations, SMEs, startups, public administrations, and private as well as institutional investors. We have also supported a variety of implementation projects ranging from 6 weeks to 9 months focused on Conversational AI (Chatbots/Voicebots), Automatic Speech Recognition, Document Processing and Computer Vision.
