This article was orginally published in Dutch by Data News. Click here to read the original publication.
As generative AI has swept the global stage, most companies are still struggling to extract its business value. This is explored in our research, where 70 percent of companies fear missing the AI race. What is hampering generative AI projects and what does it take to put it into practice?
While AI isn’t new as a technology, the adoption rate of it certainly is. Previously, companies needed huge budgets to build and train an AI model, whereas today everyone can easily use a variety of models. ChatGPT, in particular, has made generative AI widely accessible.
The generative AI hype has also reached the boardroom. However, it’s surrounded by some nervousness. Research by Cognizant shows that 70 percent of companies are afraid of missing the AI race or fear that the competition will overtake them. No wonder; in the AI context it’s not unlikely that a small company will disrupt a market where historically only large companies have been in charge.
Utilizing generative AI in an organization
Today, companies mainly rely on different pilot projects, with an expected large-scale adoption between 2026 and 2030. What’s the dilemma then? We are gradually seeing a gap between what the management wants to achieve with generative AI and what is feasible within the organization. Why is that and how can companies overcome the obstacles?
Based on my experience from various projects, these are the common risks and my recommendations on how to address them:
- Classic IT methodology falls short. Quite commonly, the necessary knowledge is not available within company walls. Since AI is about human science more than technology or data, you need to invest in people with diverse talent, including AI engineers, strategists, designers, and compliance experts.
- Difficulty transitioning from pilot projects. We see quite a few pilots, but fewer projects that succeed in taking the next step towards an enterprise-grade solution. My advice is to adopt a slightly longer horizon in the context of generative AI together with tailored methodologies that integrate technical, ethical, and strategic considerations. But don’t forget, the basics remain the most important condition for success: establish robust data governance practices before scaling AI applications at all.
- Risk of developing non-compliant solutions. Incorporate compliance and ethics from the beginning to align AI initiatives with company values. It might also be easier to start with a solution like Copilot, integrated within existing solutions in the Microsoft ecosystem. We see a lot of standard solutions appearing such as Copilot within M365, or Copilot aimed at specific functions, such as sales. You can also build a customized version with Copilot Studio, tailored to your organization.
- Lack of clarity on impactful use cases. What is the AI case where a specific company can make a difference in its sector? The entire AI process stands or falls with that foundational insight. Look beyond productivity gains only to harvest generative AI opportunities; focus on AI applications that enhance customer and employee experiences while addressing critical business processes. The more important that process is, the greater the opportunity. Generative AI can undoubtedly play an important role in a customer’s banking experience or in gaining insight, such as when collecting and analyzing logs, for example through a supply chain.
Generative AI offers organizations new ways to earn revenue, improve operational efficiency, innovate their offerings and redefine their businesses. Putting generative AI into practice, however, requires a thorough rethinking of skills, processes and structures beyond technology.
Learn more by visiting our Data & AI section of the web.
