Generative AI for businesses: Why adoption is a major challenge
Florian Piaszyk-Hensen
Director fme Products
Generative artificial intelligence (GenAI) promises significant efficiency gains and innovation opportunities for companies. However, its implementation often proves to be more complex and lengthier than initially assumed.
In this article, let us dive into practical insights from customer projects and the specific challenges faced by different departments like as legal, IT, finance, line of business, and work council. We also show which strategies and tools companies can use to ensure safe and efficient implementation.
Legal requirements of GenAI adoption
Companies must ensure that the introduction and use of GenAI is in line with applicable laws, compliance requirements, and internal regulations. In particular, the legal department should consider the following risks:
- Data protection violations: GenAI often processes personal or sensitive information. If customer data, employee data, or other protected information is fed into the model, this may violate the General Data Protection Regulation (GDPR). Companies must ensure that data is processed in a legally compliant manner and is adequately protected.
- Compliance with regulatory requirements (e.g. EU AI Act): The EU AI Act is the first comprehensive law to regulate AI in the EU and sets clear requirements for the use of AI technologies – e.g. in terms of transparency, documentation, and risk management obligations. Companies must ensure their AI systems comply with these requirements.
- Infringement of copyrights: AI models are often trained with copyrighted content. This carries the risk that generated content may contain protected works without this being recognizable. Companies must take measures to avoid unintentional infringements.
- Damage to reputation & compliance violations: Uncontrolled use of GenAI can produce problematic content, such as discriminatory, offensive, or false statements. To minimize legal and reputational risks, clear guidelines and effective review mechanisms are needed.
Challenges of different business departments in terms of GenAI implementation
Security challenges for the IT team
The IT department is challenged with ensuring the secure and efficient integration of GenAI into existing systems as well as implementing new security and data protection measures. In order to keep pace with rapid technological progress, we have identified three key areas of concern:
- Security risks: GenAI models can unintentionally disclose confidential information, especially if they are trained or used with sensitive company data. To counteract this risk, technical measures must be implemented to reliably prevent an unwanted outflow of information.
- Data protection risks: Many AI solutions are offered as public cloud services (SaaS) (e.g. ChatGPT, Gemini, Copilot). This means that company data is transferred to third-party providers, which carries potential risks. It must be ensured that this data is protected and not used to train new models.
- Skills shortage & lack of AI expertise: Many IT departments specialize in traditional IT infrastructure, networks and software development, but not in machine learning (ML) or the development of AI applications. If the IT department is already working at full capacity, the introduction of Gen AI can lead to overloading and affect other important projects. Without a clear strategy for expanding resources or the use of resource-efficient solutions, AI projects are often at risk from the outset.
Investing in AI – Finance’s perspective
For the finance department, a well-founded cost-benefit analysis is crucial to the introduction of GenAI. Only if the economic added value is clearly recognizable and long-term costs remain controllable, finance departments will give the green light for the use of this technology. The main challenges are:
- High implementation & operating costs: Powerful AI models (e.g. GPT-4, Gemini) do not come free of charge. Companies either must pay high usage fees or build their own infrastructure. In addition, there are extra costs for training, process adjustments, developing guidelines, and implementing audit mechanisms – all of which require time and resources (check the article “The Hidden Costs Of Implementing AI In Enterprise”).
- Unclear productivity gains: The use of GenAI does not automatically reduce the need for personnel, as new tasks arise, for example in the areas of quality control, error correction, and application maintenance. This can reduce the hoped-for efficiency gains. In addition, the long-term return on investment (ROI) remains uncertain because productivity gains are difficult to measure and predict.
- High risk of misinvestment: The rapid development in the field of AI carries the risk that today’s solutions will quickly become obsolete and cause high follow-up costs. Choosing a particular provider can also lead to dependency (“lock-in effect”). Regulatory changes could also result in additional compliance costs.
| Financial factor | Short-term cost | Long-term benefit |
|---|---|---|
| Implementation costs | High (infrastructure, training, audits) | Decreases as AI is integrated |
| Operational expenses | Ongoing (licenses, compute power) | Lower per-task cost over time |
| ROI uncertainty | Hard to measure at first | Becomes clearer with KPIs & tracking |
| Regulatory compliance | Expensive upfront due to legal requirements | Protects from future legal risks/fines |
| Competitiveness | No immediate edge | Strategic advantage if AI is well-implemented |
Innovation vs. skepticism of departments
While management views GenAI primarily as a driver of efficiency, many departments are hesitant to adopt it. After all, they are the ones who actually use AI in their day-to-day work – so the technology must support their actual needs to be perceived as a valuable support and not as an additional burden. Among the most common concerns are:
- Poor quality of results: AI often generates plausible-sounding answers that are factually incorrect or contradictory. If such texts appear professional but are incorrect, this results in increased correction work.
- Lack of company context: Standard AI models are not tailored to industry-specific content or the individual company context. As a result, the content generated often does not meet the actual requirements and can contain incorrect or superficial information. This also increases the risk of so-called hallucinations.
- Limited adaptability: Customizing AI to specific business needs often requires significant investment. Instead of the AI adapting to workflows, departments often must adapt to the technology, which can affect efficiency and acceptance.
- Lack of expertise & training: GenAI is often introduced in departments without adequate training. Without targeted training or intuitive tools, many employees feel overwhelmed, which increases skepticism about the new technology.
The works council’s concerns
The implementation of GenAI can substantially change the organization of work – from new workflows to job security issues. Therefore, the works council has a right of co-determination in accordance with § 87 of the German Works Constitution Act (BetrVG) and raises legitimate concerns. These include:
- Job security: The works council could fear that jobs in areas with a high potential for automation might be replaced due to AI adoption.
- Change in working conditions & equal opportunities: AI can significantly change the daily work routine and create advantages for employees who are tech-savvy. However, new requirements must not lead to employees being overwhelmed or falling behind (check articles “Improving working conditions using Artificial Intelligence” and “Generative AI and the Future of Inequality” for more details).
- Clarification of responsibilities in the event of incorrect AI decisions: Who is liable if the AI makes mistakes or generates false information (“hallucinations”)? Are employees held responsible for the AI decisions? Clear rules are needed here to ensure long-term fair and safe use of the technology in advantage of all employees.
Conclusion
GenAI offers enormous potential for companies but requires a comprehensive and holistic strategic approach. Each department faces specific challenges, ranging from data protection and compliance issues to economic risks and technical requirements. Employee involvement is essential to ensure acceptance and efficient use.
Successful and rapid implementation will only be possible if all departments work together on a well-thought-out AI strategy. This includes clear guidelines, investment in training, and the adaptation of existing processes. Companies that carefully consider these aspects can fully exploit the advantages of GenAI and ensure long-term competitive advantage.
Still unsure how to integrate AI into your business?
With GenieHub we have developed a solution to help companies use GenAI effectively and embed it sustainably into their organizations. The platform is designed for real-world challenges faced by different departments and can securely be used without any technical barriers for teams.
Want to learn more? Feel free to schedule a meeting with our experts.
Whether you’re just exploring initial use cases or shaping a full-blown AI strategy – we’re here to support you every step of the way!
