Despite the immense interest and media buzz surrounding artificial intelligence in business, its actual adoption within organisations remains limited – only around 6% of Polish enterprises use it, compared to an EU average of approximately 13% (1). This is a significant gap when measured against the ambitions and declarations of managers, who increasingly view AI in business decisions as a key driver of growth, new opportunities and competitive advantage.
In this article, based on a presentation by Michał Krupiński during a webinar co‑organised with Grant Thornton Poland, we explore how the development of artificial intelligence is transforming the digital workplace, the most common applications, where the pitfalls lie and which AI solutions are worth considering to effectively optimise business processes and enhance organisational competitiveness.
The use of AI in business processes
Looking at the data on artificial intelligence adoption, it is clear that businesses are still getting to grips with this technology. This is largely due to differences between industries and a low awareness of what AI can actually enable for organisations. In sectors such as finance, telecommunications and IT, companies tend to be more open to harnessing the potential of artificial intelligence and actively experiment with its applications.
Conversely, in areas such as construction, manufacturing or small and medium‑sized enterprises, the prevailing belief is often that AI in business is unnecessary. In reality, this stems from a lack of understanding of artificial intelligence technology, coupled with cultural and organisational concerns, leaving companies behind, even when they already have access to the tools.
This disparity is well illustrated by data showing the scale of artificial intelligence adoption among European companies. In some countries, the proportion of organisations making use of AI’s capabilities is several times higher than in Poland, highlighting how varied the approaches to these tools and to business process optimisation can be.
Companies using artificial intelligence technologies – the use of AI tools (1)
Most common applications of AI
- Text analysis
- Generation of written or spoken text
- Spoken language analysis
- Machine learning
- Workflow automation and decision‑making support
Companies primarily use AI in customer service – through chatbots, call centre systems and tools for analysing customer feedback and opinions. Such solutions support employees in responding to queries, automate processes related to returns or complaints, and enable faster processing of large volumes of text data, for example in emails or forms.
However, more advanced applications of artificial intelligence technologies, such as workflow automation or support for strategic decision‑making, are still less common, despite these areas holding significant, untapped potential.
Discover our approach to automation and AI
According to data from Harvard Business Review, as many as 80% of business leaders believe that AI will be a crucial factor for organisational success over the next two years (2). This highlights how highly managers value the potential of this technology in building competitive advantage.
In practice, however, only around 20% of companies have actually implemented artificial intelligence in their day‑to‑day operations (2). This discrepancy clearly shows that although awareness of AI’s potential is high, organisations often lack a clear strategy, the right competencies, or the courage to turn declarations into concrete actions.
The practical use of artificial intelligence in business and the challenges around AI adoption stem more from organisational culture and companies’ preparedness for change than from the technology itself.
Mistakes in implementing AI
Studies show that approximately 80% of AI implementations fail to achieve their intended business objectives or are abandoned before full deployment (3).
Importantly, the root causes of these failures are not solely technological limitations or lack of resources.
Instead of focusing on the strategic application of artificial intelligence in business, many organisations make fundamental mistakes, with the most common reasons for failure being organisational and cultural in nature:
- Unclear or poorly understood problem
Businesses often fail to clearly define the problem they aim to solve – AI is implemented simply “because everyone else is doing it,” rather than because it actually addresses a specific business need. The presence of an AI tool creates the illusion that any challenge can be tackled, even if, in reality, it makes neither economic nor organisational sense.
- Lack of appropriate data
Artificial intelligence relies on data and lots of it, at high quality. Many organisations lack structured, well-documented, and easily accessible data that can be used in a project. Without this “fuel” for AI models, systems cannot function properly and may produce erroneous results, undermining trust in the entire solution.
- Technology exceeds the actual needs
Organisations often implement technology that is far too advanced for the real-world problems they are trying to solve. Instead of opting for simple and effective solutions tailored to the scale and nature of their operations, they choose complex systems that they later fail to use effectively. As a result, investment in artificial intelligence solutions fails to deliver the expected outcomes and may even complicate existing processes.
- Problems with infrastructure and implementation
Deploying AI-based tools requires the right technical infrastructure and a team with appropriate competencies. Many organisations lack both IT resources and personnel capable of implementing and maintaining such solutions, or of understanding the specific nature of AI – such as an AI researcher, who can assess technological capabilities and align solutions with business needs.
- A task too difficult for today’s AI
Sometimes companies expect more from AI than current technology can deliver. They set overly ambitious goals, such as full automation of complex processes, and are disappointed when the system falls short. However, it is worth noting that this is by no means the most common cause of failure – far more often, the problem lies within the organisation and how it has prepared for the implementation.
According to data, only around one in five digital transformation initiatives achieves its intended objectives, with an unsupportive organisational culture being the common denominator of failures (4).
Artificial intelligence in business
Despite the growing role of artificial intelligence tools, their potential benefits for businesses, and increasing interest in technologies transforming the digital workplace, actual usage among employees remains relatively low.
Data show that as many as 81% of employees do not use generative artificial intelligence tools in their daily work, and only 16% report any experience with such solutions. This figure may even be underestimated.
Such a level of adoption largely stems from varied attitudes towards the technology itself. Some employees react with denial – convinced that AI cannot replace their skills, they dismiss its potential and avoid engaging with new tools. Others feel uncertainty and fear, they have limited understanding of how AI systems work and worry they might lose their job or see their role diminished.
Artificial Intelligence in everyday work
In practice, many employees admit unofficially that despite no formal rollout, they integrate AI tools like ChatGPT or Google Gemini into their digital workspace. They log in with personal accounts and use them as an alternative to traditional search engines, leveraging AI primarily to look up information or generate text: asking it to compose emails, refine tone, or draft brief summaries.
While these practices often enhance efficiency and communication quality, employees tend to conceal them from managers due to unclear guidelines, fear of negative judgment, or concerns about compliance with company policy. This highlights that effective and strategic implementation of artificial intelligence requires not only appropriate tools but, crucially, education, transparency, and trust within the workplace.
Market data analysis
A study on a representative sample of Americans shows that usage patterns vary considerably depending on the industry. (5)
Communication and training – the missing links in AI adoption
Data show that only 15% of employees state that their organisation has clearly communicated its strategy and plan for AI implementation (5). This means that in the vast majority of cases, employees do not understand the purpose of the implementation or how the technology is supposed to support their daily work.
Moreover, even in organisations where AI is already in use, employees rarely receive adequate support. Only 47% of people in companies that use generative artificial intelligence confirm that they have been provided with any form of training in this area.
This lack of communication and education deepens the AI skills gap – the mismatch between what is expected of employees and their actual skills – and leads to uncertainty, resistance to change, and underutilisation of the tools’ potential. If organisations want to consciously shape the future of artificial intelligence in the workplace, they must ensure transparency, education and employee engagement.
Explore our training on artificial intelligence – AI in research team workflows.
The potential of Artificial Intelligence as a competitive advantage
A lack of communication and training means that employees do not know how, or why to use new tools. This is particularly evident in the case of internal chatbots, which companies often develop with an overly broad range of functions – text generation, content creation, document analysis, answering questions, reporting, and more.
Such an “all-in-one” product often proves to be of little use in practice. Users do not understand its specific purpose and limit themselves to the simplest commands, such as writing an email or editing a text. The rest of the functionality remains unused.
Much more effective are narrowly focused products, designed to solve a specific problem. An example might be a chatbot dedicated solely to managing travel expense claims: it automatically reads data from invoices and fills in the forms. Tools like these are clear to users and far more likely to be embraced in everyday work.
Artificial Intelligence in business – how to implement it?
Successfully implementing artificial intelligence in business requires a thoughtful, step-by-step approach. Crucially, the process should not begin with the choice of technology but rather with understanding the organisation’s real needs and identifying the problems that artificial intelligence encompasses and can help solve.
Conducting a process audit and identifying issues allows organisations to pinpoint areas where automation or data analysis can genuinely deliver time savings, cost reduction, or improved quality of work. Before choosing an AI tool, it is essential to thoroughly analyse existing processes and identify bottlenecks, costs, and employees’ pain points.
- Where do we lose the most time?
- Which tasks are repetitive, tedious, or error-prone?
The aim is to identify problems which, if solved, would bring tangible improvements to the organisation – such as faster information flow, reduced manual effort, or better communication.
The next step is to assess AI’s capability to address the specific problem. Once we know what we want to improve and where to introduce process automation, we need to honestly evaluate whether artificial intelligence encompasses a solution to the issue.
Not every business challenge requires artificial intelligence – sometimes better data structuring or improved procedures are sufficient. However, if AI is indeed the right choice, it is crucial to ensure the data that fuels it is of high quality, complete, and readily accessible.
It is equally important to prepare resources properly: data, infrastructure and tools, the team, and the necessary skills.
The best practice is to start with small, pilot MVP (minimum viable product) projects, which can be quickly tested in a real working environment and gradually developed.
This approach allows assumptions to be validated against reality and enables early feedback to be gathered from users. Successful projects show that, in the strategic application of artificial intelligence, it is worth focusing on a few key use cases with the highest potential and directing efforts there. Once the prototype is implemented, measure the results.
Frequently Asked Questions (FAQ) – business process optimisation
Do ethnographic research and shadowing help when implementing AI in business?
Yes. Ethnographic research – observing how employees actually use artificial intelligence tools and systems in their daily tasks, helps uncover problems that do not emerge in surveys or interviews. Employees’ declarations often differ from reality, and such research enables the design of solutions that are better suited to their real needs.
Is it worth starting with an MVP when purchasing a new system?
Yes, the MVP (Minimum Viable Product) approach is not just a passing trend but a proven practice. A small, pilot deployment allows you to test the solution in a real-world environment, engage users, and avoid mistakes during a larger investment. MVP also helps foster a positive attitude towards change, identify potential issues early on, and increase the competitiveness of firms that adapt to market needs more quickly.
Why do employees hide their use of AI?
They often fear their managers’ reactions, are unsure whether using private AI tools, such as ChatGPT is acceptable, or worry they might be perceived as less engaged or even lazy. The lack of clear guidelines and communication within organisations deepens this uncertainty. However, an open discussion about this phenomenon and establishing transparent rules are crucial steps towards building a culture of trust and preparing organisations for the future of artificial intelligence in business.
How can resistance to implementing AI in business be overcome?
Key actions include clear communication of objectives, training tailored to employees’ knowledge levels, and involving staff in the process and business decisions from the very start. Change ambassadors, who test solutions and support their colleagues in daily use, can also play a vital role.
Does automation and implementing AI always require advanced technology?
No. Many problems can be addressed with simple, specialised tools that do not require complex infrastructure but still offer new possibilities for artificial intelligence. It is essential to match the technology to the real needs and capabilities of the organisation, which can be identified, among other methods, through market data analysis. Such analysis helps to better plan the use of artificial intelligence in optimising business processes.
