GenAI: Strengthening the Business Vision Behind the Innovation

Generative artificial intelligence (GenAI) solutions, as widely documented, can experience what are known as “hallucinations.” In response to a question or instruction (known as prompts in the AI space), they can return absurd, false, or biased information. Surprisingly, experts in artificial intelligence are not particularly alarmed by this issue.

According to specialists, in the most robust GenAI tools, these digital delusions appear in less than 20% of prompts. Moreover, as AI technologies are exposed to increasingly complex topics, it is expected that these hallucinations may occasionally occur — after all, GenAI requires time and data to refine its understanding of sophisticated subjects. This learning process will inevitably involve some errors and missteps.

But beyond these digital hallucinations, a different kind of delusion may emerge, one that takes hold of an organization dazzled by the technology's capabilities and eager to join the AI trend. This excitement, however, does not guarantee a happy ending.

The Challenge: GenAI With a Business Perspective

In the corporate environment, recent research shows that 95% of GenAI initiatives fail. At the root of this alarming trend is a common mistake: companies attempt to replicate, within the enterprise, the same capabilities available in popular tools like ChatGPT, one of the most widely used GenAI applications in the world.

However, mimicking features, even those from the most celebrated platform, is not a viable path for business. For GenAI to work, a company needs a strategic implementation aligned with its business characteristics and needs. This includes understanding the challenges and risks associated with deploying such innovation. A company cannot apply the same logic used in mass-market, general-purpose applications.

Moving away from popular references like ChatGPT leads to a critical first step: deeply studying the technology to understand its core processes, implementation models, operational challenges, infrastructure requirements, management mechanisms, and relevant case studies. This learning process is essential for developing a GenAI solution that truly meets business needs.

With this deeper understanding, companies will be better positioned to design GenAI solutions that are fit for purpose: identifying the business areas and activities that will benefit most, recognizing the risks involved, and applying best practices to unlock the technology’s potential.

From this perspective, organizations can take several strategic measures:

Apply GenAI in the right place an under the right conditions. A company with a clear vision of GenAI’s current capabilities will understand that this technology at its present stage delivers the most immediate impact in areas with structured data and well-organized processes. A great example is customer service departments, where operational manuals already exist for client interactions. These manuals can be used to train a GenAI agent to deliver support that is fully aligned with the company’s protocols and policies. Conversely, areas with inconsistent processes or data should not be top priorities for GenAI implementation.

Reinforce the importance of corporate data. As with all branches of AI, generative tools require high-quality, reliable data to deliver optimal results. Without it, the risk of hallucinations and poor outputs increases significantly. Implementing GenAI should therefore serve as an opportunity to strengthen data governance policies and practices. If a company cannot be confident in the integrity of its information, it must seriously reconsider whether it is ready to adopt GenAI.

Address cybersecurity concerns. GenAI solutions have also caught the attention of cybercriminals. A security-conscious organization will take proactive steps to protect its GenAI environment; for instance, securing training data and processes to prevent malicious actors from manipulating the AI’s learning, and implementing safeguards to ensure that generated content (text, audio, video, graphics) does not expose confidential company information. This also involves setting specific rules: What corporate data should GenAI be allowed to access and use?

Define ethical guidelines for GenAI usage. A company with a serious approach to GenAI will not only comply with relevant regulatory frameworks, but also acknowledge the non-negotiable ethical dimension of this technology. GenAI-generated content can, for various reasons, be false, harmful, or offensive. A well-prepared organization establishes governance mechanisms and policies to prevent such outcomes. It promotes the use of governed, socially responsible, and ethical AI.

Without a doubt, generative AI will have a significant impact on the business world. But to harness its full potential, companies must approach it with strategic insight and informed decision-making. That potential — the kind that delivers real business outcomes — won’t come from simply using a popular app. It must be shaped and developed through deep understanding, proper alignment, and enterprise-grade implementation.