Navigating the AI Hype: Best Practices from Jakub Michalička

Jakub, we're marking two years since the public release of ChatGPT. How would you summarize the journey so far? 

It’s been fascinating to see the recent progress across various AI fields. While generative models like ChatGPT and neural networks have captured much of the attention, significant advancements are also being made in supporting areas such as AI chips, hardware development, and tools for AI professionals. The intense competition among big tech companies in this emerging market drives innovation, and we benefit from the resulting products and solutions. 

In the past two years, AI has evolved from a niche technology to a widely discussed topic – perhaps even overly so. We’re now entering what Gartner calls the "Trough of Disillusionment," a natural phase in every hype cycle. 

What have you observed about how businesses adopt AI over the past two years? 

AI is just a tool – a powerful one, but a tool, nonetheless. Its success depends on how well it’s aligned with solving specific business problems. Unfortunately, early on, there was a significant "FOMO effect" (fear of missing out). Many companies rushed to implement AI without understanding what it was meant to accomplish. Unsurprisingly, this often led to disappointment. 

Recently, we’ve seen a shift. Companies are becoming more knowledgeable, and aware of AI’s limitations and potential pitfalls. In some cases, organizations are even overly sceptical or hesitant, often due to previous negative experiences or fatigue from AI's omnipresence in conversations. 

The key takeaway is that companies must start by clearly defining the problems they aim to solve and then evaluate if AI is the right tool. A well-planned, use-case-specific implementation is far more likely to succeed than rushing into AI developments without much prior thought. 

What key trends are you seeing in AI today, and where do you foresee future developments? 

Predicting the long-term trajectory of AI is challenging because it evolves very quickly. Industry experts generally anticipate a 30-40% annual growth rate in the AI sector over the next five years. 

From an enterprise perspective, I see two main directions for future developments: 

1. AI for Every Office Employee

Companies like Microsoft and Google are integrating AI tools into their platforms to streamline everyday tasks, such as summarizing emails, generating meeting notes, and preparing presentations. These so-called Copilots or agents aim to boost productivity. Early versions of these tools had their flaws, but improvements are coming fast.

If you’ve tried these tools before and were disappointed, I encourage you to give them another shot – they’re becoming surprisingly effective. 

2. Specialized AI Solutions

On the other end of the spectrum, we’re seeing the rise of AI tools tailored to specific industries. For example, healthcare is leveraging AI for diagnostics, while supply chain companies use it for predictive analytics. Navigating the crowded landscape of AI startups and solutions can be daunting but investing the time to identify tools that genuinely work pays off in the long term. 

Now to the trends-related question. Three major topics are shaping the AI conversation across these areas: 

• Explainability: Many AI systems function as “black boxes,” meaning their internal workings are not fully understood even by experts. This lack of transparency can undermine trust and make businesses hesitant to rely on AI without human oversight. 
• Ethicality: Large Language Models, but also other AI solutions, can unintentionally reinforce biases or mishandle sensitive data, leading to ethical dilemmas. Businesses must be vigilant about these risks and be proactive in preventing such issues. 
• Efficiency: AI can be resource-intensive, requiring significant energy and infrastructure. The industry is addressing this with advancements in hardware, downsized open-source models, and a growing emphasis on using AI responsibly only when it adds value. 

How can Actum Digital help businesses implement AI successfully? 

Implementing AI requires careful consideration of several key factors: 

Business Case: Does your AI initiative align with your overall strategy? Is AI truly the best tool for the problem, or is there a simpler solution? And if AI is the right choice, is there a clear return on investment? Sometimes, off-the-shelf solutions may meet your needs, while other times, custom development is required. 

Technical Preparedness: Is your organization ready from a technical standpoint? AI requires high-quality data, robust security, and scalable IT infrastructure. You also need a vision for ongoing maintenance and operations. 

Change Management: Technology adoption isn’t just about implementation – it’s about ensuring people use it correctly and effectively. Businesses need plans for training employees, integrating AI into workflows, and aligning usage with ethical and organizational standards.

And be prepared for the possibility that AI solutions might not work initially. This is especially true for custom solutions, which often involve experimentation and early failures. Conducting a feasibility study upfront is essential to evaluate the model's performance and identify potential challenges early on. Be ready to test different approaches and refine your strategy before deciding on the best path forward. 

At Actum, we support clients through every stage of their AI journey. From ideation and feasibility studies to development, implementation, and integration, our team of experts ensures that businesses unlock the full potential of AI while mitigating risks.