Problems and challenges

Many companies, projects, and startups are focusing on developing generative AI tools. However, it’s essential to acknowledge that AI is merely the final step in a process that begins with datafication and the verification of facts. While generative AI can help convert unstructured data into useful information, it’s much better to structure and secure the data first. Ensuring that you use a reliable dataset for your prompts is crucial; otherwise, you risk a scenario where you have garbage in and potentially even more garbage out.

On a global scale, we produce at least 20-40 MB of data per person per minute, but only 5-10% of this data is structured and trustworthy. Imagine the possibilities if we could trust the data we rely on. What if generative AI utilized reliable data?

Superior data quality significantly enhances the autonomy of AI agents, empowering them to perform more effectively and independently.

There are two types of problems: urgent and important. Urgent problems are not necessarily important, while important problems are rarely urgent. Urgent issues focus on improving existing solutions (continuous improvement and sustaining innovation), whereas important issues aim to create something entirely new (discontinuous improvement and disruptive innovation). Our focus is on addressing important problems. To achieve this, we need to adopt a new innovation approach.

Developments
Given the poly-crisis in which we find ourselves, humanity has arrived at a point where we are rethinking the starting points from which we want to shape our society. Thinking about how we organize our work and make our prosperity more sustainable. The next step could be transitioning to a sustainable, data-driven, decentralized, and more humane society and economy. A society with far fewer local IT systems in which institutions and technology are not leading but serving.

The transformation is necessary and in full swing. What role does your educational curriculum, research program, or organization play in this? Increasingly, organizations are becoming data-driven. No longer reasoning from the company’s perspective but from new entrepreneurship, people, and data. There is no local HR, CRM, or accounting system, but AI agents that communicate with reliable datasets. We have the technology to simplify a complex world. We have the technology to create a shared, secure, and accessible reality. No longer taking IT as a starting point, but data. Even hyped trends like machine learning and artificial intelligence cannot do without rich data. But what exactly is data? And what are the characteristics of rich data, and how do you organize the production and consumption of data?

We all tend to keep reality in our local IT systems, but this is a reality of our own and often not a shared, secured, and accessible reality. The fact that, despite billions of investments, we hardly become more productive and that we usually spend more than a third of our time requesting data from other parties says it all. In any case, you must ensure that agreed facts are stored in such a way that they cannot be manipulated unilaterally. Then you organize access to data. So, we are no longer pursuing data ownership, but rather organizing access to data. The GDPR is very clear about this with data minimization.

The future of every organization depends on the answer to the question: how data-wise or mature are you, and is your organization data-driven, sustainable, and decentralized? How do you ensure that unnecessary processes in supply chains disappear and that necessary processes run more safely and efficiently? Not always by listening to your IT department, existing software supplier, or consultancy firm. They may have other interests, rooted in old factory DNA, and take too little account of the relative simplicity of data and data organization.

We all know the saying: “You can make anything with Lego”. The same goes for data. What electrification meant to factories, datafication means to offices. What the physical assembly line is to factories, the digital assembly line is to offices. But how will you set up and utilize the digital assembly line within your organization? How will you let go of your dominant logic and organize it fundamentally differently? What are the consequences for your operating model or educational curriculum? Before making radical changes with high investments, it is wise to make future systems transparent through prototyping and simulations, so that people understand, trust, and embrace them.

Challenges for your organization
The organization of data and information, including AI, faces several significant challenges. Many individuals no longer recognize the value of their work, and three out of ten people experience information overload. Despite substantial investments in IT, productivity is growing at a barely noticeable rate, while the costs to our prosperity continue to rise by 2-3% each year. We require a 1.5% productivity growth to offset the impact of an aging population alone.

Additionally, cyber attacks are almost impossible to defend against, and we contend with issues such as biased algorithms, propaganda, fake news, disinformation, and AI-induced hallucinations. The current business models of major tech companies are unlikely to address these problems.

Furthermore, the privacy and protection of sensitive company data cannot be adequately guaranteed. Only 8% of companies are considered data-mature, and a staggering 88% of data integration projects and 70% of digital transformation initiatives fail to meet their objectives. According to the Boston Consulting Group, only 30% of digital transformation projects are successful. Less than half of organizations express confidence in the quality of their data, with only 15% expressing confidence in the quality of their external data. Alarmingly, 25% of the content in CRM systems becomes outdated within a year. How is that possible if we spend billions of dollars on IT worldwide? The problem is not IT, but the way we organize data, office work, trust and attention.

Waste is the death of any achievement.

However, if we organize data more intelligently, we can operate with fewer data centers and AI trainers, which would be more environmentally friendly. Lastly, we waste over 30% of our valuable time on unnecessary computer tasks, translating to a lot of  digital waste. Gartner has found that the average cost of poor data quality on businesses amounts to anywhere between $9.7 million and $14.2 million annually. At the macro level, bad data is estimated to cost the US more than $3 trillion per year (Harvard Business Review). In other words, bad data is bad for business or to train and use AI. These figures are comparable for Europe and the Netherlands. We could save 2,3 billion hours or 1.8 million FTE in the Netherlands. More than enough to absorb the consequences of an ageing population, for example. The problem is that we don’t have a shared, trusted and accessible reality despite billions of investments. Not the technology, but the way we organize supply and demand of data is the problem. Organizations worldwide are looking for effective ways to reduce digital waste, lower information costs and improve customer services. Generative AI as part of the digital assembly line is the answer. This presents a significant opportunity for system innovation.

AI developments and challenges in the workforce
Recent studies* highlight several key findings about the state of AI in the workplace:

  • Employer engagement: 84% of employers and HR managers do not consider AI when performing their tasks. Furthermore, 40% are not yet using AI, and half of those have no plans to implement AI in the future;
  • Public usage: 58% of Dutch individuals use AI personally or professionally, but significant concerns remain;
  • Concerns about AI: 85% of respondents cite negative effects of AI as their biggest concern, which include inaccurate results, loss of control, privacy violations, and the risk of disinformation.;
  • Trust in information: 72% of people are uncertain about the reliability of online information, a concern exacerbated by AI applications like deepfakes and automatically generated content.
  • Need for regulation: 82% believe they would have greater confidence in AI if better regulations were in place, and 85% advocate for legislation to combat AI-generated disinformation;
  • Awareness of existing regulations: only 11% of respondents are aware that regulations, such as the AI Act and AI literacy requirements, already exist.
  • Strategic priorities: 65% of senior executives prioritize AI as part of their strategic initiatives;
  • Informal Deployment: 27% of companies are using AI informally across various teams without a comprehensive strategy. The main challenges faced include:
    • Privacy and security concerns (38%)
    • Lack of budget or specialized staff (28%)
    • Insufficient data quality or standardization (26%)
    • Redundant and unstructured data sources (25%)
    • Absence of a central data strategy (24%)
  • Data fragmentation: 76% of professionals indicate that data fragmentation hampers their ability to obtain real-time information;
  • Budget allocation for transformation: 30% of senior executives plan to increase their technology and data transformation budgets by over 10% by 2025 to improve data organizational practices prior to fully committing to AI-driven decision-making.

In summary, the development and utilization of AI face significant obstacles, including data silos, a lack of transparency, and the challenge of advancing projects beyond experimental stages. As a result, companies are increasingly recognizing the importance of AI literacy and data-driven organizing within their value chains. Without a clear vision, a proper evaluation of opportunities and risks, and a robust data foundation, AI can be not only ineffective but also quite risky.

* See for instances:
1. The age of Intelligence, A perspective on Trust, attitudes and use of artificial intelligence: A global study 2025 (KPMG International, The University of Melbourne).
2. Adobe 2025 AI and Digital trends report.

Responsible AI
Weconet develops and utilizes generative AI consciously and responsibly to genuinely enhance productivity and create surpluses that sustain our prosperity. Therefore, we also focus on the following aspects of generative AI:

  • privacy, ethics, and morality
  • (data) security & quality, rich datasets
  • governance, policy, and legislation
  • culture, norms, and values
  • private/public and open source
  • perspectives and interests (‘cui bono’)
  • biased discriminating algorithms
  • energy and water consumption
  • rebound effect (Jevons-paradox)
  • productivity paradox

Get in touch for more details.

tags: problem, challenge