Without Ample High-Quality Proprietary Data, AI Simply Cannot Deliver, Asserts David Kaláb of Adastra

April 15, 2024

In the realm of artificial intelligence, groundbreaking innovations are likely behind us. Instead, companies must prepare for a sobering reality check within the next two to three years. They’ll discover that the effectiveness of artificial intelligence (AI) hinges on the quality of data. Without ample relevant proprietary data, the promised business benefits remain elusive. David Kaláb, Vice President for Data Management at Adastra, highlights this in an article for CzechCrunch, offering five essential tips for navigating the data landscape to ensure your business stays on track in the coming years.

Success with artificial intelligence isn’t just about adopting new technology; it’s also about a strategic approach to data and its management. It’s high time to invest in data management, security, and cloud technologies to ensure companies are innovation-ready. How should we view AI and data, and how can we utilize them to secure the future prosperity of a company?

1. Prioritize proprietary data

For real AI success, access to unique, proprietary data is essential. They provide deeper insights and enable tailored solutions for your company, something that market or competitor data generally can’t offer. Investing in the collection and analysis of proprietary data will become your competitive advantage.

2. Data governance is indispensable

Good data governance ensures you’re working with quality, secure data and in compliance with legal standards. Effective data management is the foundation for reliable and ethical AI utilization.

3. Data security above all

Data security must be a key part of your AI strategy. Once data leaves your organization’s secure perimeter, it could be exposed to risks. Data security protects not only your information but also your reputation and customer trust.

4. Transition to the cloud is a necessity

Cloud solutions provide the agility, scalability, and innovation potential necessary for effective AI utilization. Transitioning to the cloud is a crucial step to ensure your company stays ahead in innovation and doesn’t lose competitiveness.

5. Prepare for the disillusionment phase

Expert analyses, including those from Gartner, predict that the AI excitement will be followed by a period of disillusionment, where companies realize that AI success requires more than just technological adoption. It’s important to have realistic expectations and a long-term strategy for integrating AI into your business.

Quality data is an essential part of artificial intelligence

Many companies have yet to realize that without data, artificial intelligence won’t propel their business forward. Data is like the chassis of a car. Companies need an adequate amount of quality data, coupled with good data governance and strong security practices. Without a solid chassis, no car can run smoothly.
One example illustrates this: imagine a large insurance company deciding to efficiently manage its building’s consumption. It needs to track how many people move where and adjust internal conditions accordingly, such as temperature, heating, ventilation, and so forth. It installs sensors in the building and collects data. But, how does the company handle the data from the sensors?

How do they interpret them and evaluate them, and according to which algorithm? How do they clean the data from the sensors? Without good data governance, organizations are missing out on crucial insights.

When artificial intelligence becomes a customer itself

All technologies undergo a similar life cycle. Initially, there’s a steep curve of enthusiasm for the novelty, but gradually, companies realize it’s not as straightforward as it seemed at first. The market sobers up, and only then do solutions and best practices emerge that actually work. This was the case with cloud, IoT, and now with artificial intelligence.

In three to five years, new products will be developed with artificial intelligence, not only in the automotive sector (where we have already observed it) but also commonly in healthcare or education.

While many solutions may be ecological, they certainly won’t be economical.

The ultimate success will be when artificial intelligence itself becomes a customer. Can you imagine a world where objects behave like customers? A car orders wiper fluid water for its windshield wipers when it’s running low. In case of a malfunction, it finds an available slot at a service center and schedules an appointment. Similarly, dishwashers or washing machines could monitor their consumption of supplies such as detergent or cleaning tablets, and restock when running low.

Use cases like these will transform the market, but not without its challenges. Setting up a flexible supply chain management to accommodate the behavior of these ‘new customers’ will be a non-trivial task.

Digital investments and best practices are lacking in the market

Based on market experience, I don’t expect this trend to emerge rapidly. It would require significant digital investments, and the creation of solid market best practices to build upon. We’re able to optimize many processes using AI, but economically, it doesn’t always pan out.

A practical example: When you order groceries from Rohlík or Košík, the delivery driver usually brings them in many paper bags. This happens because the groceries are picked by several pickers, each with their own portion of the warehouse. So, it’s common that you buy eight products and receive them in several paper bags. It’s less eco-friendly but efficient for the seller. If we wanted to optimize the picking process in the warehouse and deliver the purchase in, for example, two bags, the whole process would significantly slow down, resulting in economic losses for the seller.

Furthermore, companies will also struggle with sustainability. While many solutions may be ecological, they certainly won’t be economical. Regarding sustainability, we’re talking, for example, about the use of visual AI. If I were a bank and wanted to install cameras at branches and use artificial intelligence to improve the customer experience, it would be extremely expensive. I would need huge computational power and massive storage capacity to process visual material. Not to mention the data waste that I would need to address in this context.

The article was published on the CzechCrunch portal on March 19, 2024.