Although artificial intelligence is nothing new and has been circulating within academic walls for quite some time, it has only become widely accepted and widespread in the public domain in the 21st century.
It's hard to say that AI was ever completely irrelevant, but it was certainly limited to certain circles. Its rise to the mainstream can be attributed to several key factors: technological advancements, a change in perception about AI, increased investment in this technology, and, unavoidably, the media.
Media fascination with AI, along with often bombastic headlines about how AI will replace us all, has further spurred public curiosity and put artificial intelligence under the microscope.
In this blog post, we will touch on the impact AI and efficiency have had on today's work and, more clearly, how we use AI for worker productivity at Devōt.
How to achieve more with less - Subtractive productivity
According to the Encyclopaedia Britannica, artificial intelligence encompasses the ability of digital computers to perform tasks we typically associate with intelligent beings. It's important to note that AI encompasses a wide range of concepts and applications, but the AI that most of us use in a business environment belongs to generative AI. This form of AI, which dates back to the last century with examples like the chatbot ELIZA (long before our chatbot Ante), is today most famously implemented through applications like ChatGPT.
Thinking about employee productivity often boils down to the amount of work done during the day. However, the perception of productivity varies, especially in the business world, where employees and team leaders do not value productivity in the same way. In a business context, productivity is often viewed from a strategic perspective, looking at how individuals or teams contribute to the broader goals of the company.
Subtractive productivity shows how it's possible to achieve more with fewer resources. The idea is that productivity should be a "process of reduction": we should aim to reduce errors and unnecessary steps in our work. This approach not only increases efficiency but also leads to higher-quality work outcomes.
What is the difference between efficiency and productivity at work
Considering I am mentioning in this blog AI and efficiency, and then AI and productivity, just to be clear on vocabulary.
Efficiency refers to performing tasks correctly with minimal resource waste. Productivity, on the other hand, measures the output produced within a given timeframe.
In the workplace, AI-driven efficiency and productivity are interconnected but focus on different improvement aspects. For example, if employees use an AI tool for quicker learning and skill acquisition, the efficiency of their learning process increases—they spend less time and effort to gain necessary skills. This efficiency directly impacts productivity; as employees become proficient more quickly, they can take on more complex tasks sooner and with better competence.
So, to sum up, productivity is more about the output (employees contributing to a greater extent), and efficiency (for example, the learning curve) focuses on the input.
AI efficiency at Devōt: Different perspectives on using AI for work productivity
We keep seeing how jobs adapt and change. In the technological world, where adaptability is the only constant, artificial intelligence (AI), mainly machine learning models, plays a major role in changing business.
And no, this isn't another bombastic article about AI replacing human labor. Instead, we'll see how, through the application of generative AI, we can work on our version of "Subtractive Productivity." In our company, the use of AI varies depending on the role, but most of us started to use AI after the boom of ChatGPT.
Developer perspective - AI enables a smooth transition between learning new technologies
Our Tech Lead, Marko Meić Sidić, regularly uses AI tools like Copilot in his IDE. Copilot significantly speeds up writing boilerplate code and offers suggestions for refactoring and generating unit tests. Marko particularly highlights that AI saves him time with education or when transitioning from one programming language to another. After more than five years of working in Ruby on Rails, last year, he started programming in Java and the Spring framework, where AI significantly accelerated his adaptation and enabled a smooth transition to another language.
While AI helps speed up the process, Marko emphasizes that his approach to problem-solving remains similar because AI still requires human input. He highlights the importance of properly formulating queries in AI tools: better context and higher-quality data result in better responses. If adequate context is missing, AI can actually make the process worse, as you spend more time arguing with it than solving the problem.
Marko also adds: "I always try to come up with solutions myself, but that certainly includes a combination of official documentation and then AI. AI can definitely speed up the process, but it's always necessary to verify the solution AI proposes. Regardless of the speed of progress of AI technologies, they must continuously align with rapidly developing technological standards."
Despite AI's help, Marko emphasizes that Code Reviews by other developers and work by QA engineers are still essential. AI cannot fully replace thorough quality checks and manual testing, which are crucial for maintaining high software development standards.
QA perspective - Artificial intelligence cannot replace critical thinking
QA Leo Cvijanović uses AI in his job, especially for writing automated tests. In addition to ChatGPT, he also uses GitHub Copilot, which he finds extremely useful due to its integration into the code editor. Copilot has the ability to use existing code within the project or repository to help create new code.
Copilot allows him to simply describe what he wants to achieve, enter the necessary variables and parameters, and then the tool takes care of the syntax of the code. This automation allows him to focus more on the more complex aspects of the QA process.
Despite the advantages that AI systems offer, he believes that the QA position is still just as necessary. Regardless of the advancement of AI technologies, the code must undergo a detailed code review before reaching the QA phase, where human intervention is necessary for the final check. Leo believes that AI, no matter how advanced, will never be perfect and cannot completely replace the critical review and analysis provided by the human factor. Therefore, while AI can significantly improve the efficiency of processes, it is important to remain cautious and not rely solely on technology.
Designer perspective - AI is never the ultimate solution, but a help on the way to the solution
Our designer, Tisa Bastijanić, uses AI technologies to expand and generate images in Photoshop, while in her role as Product Owner of Devōt's website, she most often uses AI for writing User stories. Tisa points out that although AI significantly helps in the efficiency of her work, she never uses it as the final solution. For her, AI is a tool for support and quick access to information, serving as a help on the way to the solution, but not as a replacement for the creative process.
Tisa also emphasizes the importance of recognizing moments when it is better not to use AI. Due to various shortcomings, such as "hallucinating" answers or misunderstanding queries, using AI can result in spending more time than necessary, sometimes making manual work faster and more efficient.
In the world of design, AI-generated images, from portraits of human faces to scenic elements, often contain illogical things or photos look just "too perfect." So far, Tisa has not used any images or graphics created by AI in her projects. Often, illustrations are easier and faster to find via the internet than to correct AI-generated illustrations. However, she believes that the quality of AI tools will greatly improve in the future.
The perspective of a Talent Acquisition Specialist - Today, there is even greater emphasis on technical interviews
Lina Višić, a Talent Acquisition Specialist at Devōt, uses AI most to enhance the recruitment process. AI helps her in various aspects, including writing procedures, defining job descriptions, generating interview questions, and structuring meetings. She particularly highlights the usefulness of AI when she lacks ideas for new content.
Lina believes that the quality of job applications has always varied; there have always been poor and good resumes. Previously, candidates often used the first available templates from the internet, while today standards are becoming somewhat higher. However, it is easy to notice which resumes are generated without additional processing and adaptation.
A major change in the hiring process for technical staff is extra emphasis on verifying technical knowledge. Given AI's capabilities, such as ChatGPT, there is a risk that candidates will submit codes generated using AI without a deeper understanding. At Devōt, we solve this problem by conducting thorough technical interviews that test not only the accuracy of the code but also the candidate's understanding of the methods of using AI for professional purposes.
Although Lina acknowledges that AI can sometimes slow down the process due to the time required to formulate effective queries, she believes that AI has significantly helped increase the efficiency of performing business tasks that are otherwise not her favorite (and yes, we are talking about writing down some boring procedures).
So, how does AI increase efficiency in operational work?
In operational work, AI plays a key role, especially through applications of generative and predictive artificial intelligence. Although the implementation of artificial intelligence varies depending on the industry, in this blog, we focused on everyday business tasks. Our employees know that AI is a tool, not a standalone solution. We went over some of our colleagues' perspectives on AI and efficiency to sum up some key benefits that AI brings to business processes:
1. Automation of routine tasks
We all know how tedious and time-consuming repetitive tasks can be. According to research, a typical office worker spends about 10% of their working hours manually entering data into business applications like ERP and CRM systems, while more than 50% of working hours go to creating and updating documents. Automating these tasks through AI allows employees to free up time for learning and other more productive activities.
43% of the developers said that learning new skills affects their workday most positively
Automation, especially that driven by AI technologies, frees up time that developers can use to develop and refine their skills. When developers were asked what most positively affects their workday, they highlighted learning new skills (43%), receiving feedback from end-users (39%), automated tests (38%), and designing solutions for new problems (36%) as key factors.
How do you make time for all these desires during work? Automating some aspects, such as testing and code generation, allows developers to devote more time to more complex and creative aspects of their work.
According to research, 70% of development developers believe that using AI tools for coding will provide them with an advantage in work. Skill enhancement is considered the main advantage, followed by increased productivity.
This indicates a cause-and-effect relationship between increased productivity and opportunities for personal development: better productivity allows more time and resources for improvement and a focus on innovation.
2. Improved decision-making processes
AI can quickly analyze vast amounts of information, which is extremely helpful in decision-making processes. According to research, 85% of business leaders have experienced stress when making decisions. AI can help in formulating additional options, identifying key points, and enabling easier organization and access to company data. Using AI as a decision support tool can significantly reduce stress and improve the quality of decisions.
3. Easier learning (and transitioning between programming languages)
Software development can particularly benefit from AI solutions that optimize the transition between different programming languages.
As we saw in our company's example, this capability improves productivity and flexibility in AI software development and helps developers learn programming languages faster. I mean, who would have thought a few years ago that one could switch so easily between Ruby and Java?
4. Possibility of making customized generative AI
Specific business needs may require the development of customized AI solutions. Generative AI can be customized to solve specific challenges, providing support in creating unique solutions or content that directly improve operational efficiency and meet specific business needs.
In our company, in response to the need for efficient processing of large internal documentation, the chatbot Ante was created. However, it is important to note that the development of customized AI solutions is not limited to developers alone. There are tools that allow users without programming skills to create their own GPT models, further facilitating access to the benefits of AI technology within different departments.
AI is not a magical tool
Although artificial intelligence is a powerful tool (yes, it is increasing productivity, and we should leverage AI), it is not a magical solution that will perform all our tasks independently. As we have learned from our Agile colleagues from Devōt, tools are here to help our work, not to take it over completely. We should use AI to solve specific business problems, not as a replacement for human labor. While it is an opportunity for problem-solving, it is crucial to understand its limitations.
We explored all sides of AI in this blog, but let's go over a few points:
-
AI can offer irrelevant or generic answers
-
AI hallucinations and fact fabrication
-
Low quality of data on the internet
-
Excessive reliance on technology and neglect of our security
While technology should assume responsibility for repetitive and mundane tasks, freeing us to develop other skills, it raises an important question: Is there a risk that we might neglect to cultivate our own problem-solving abilities? Moreover, how secure is AI in reality?
These questions were answered at the presentation of our Business Analyst, Sandro Bujan, called "Is the fear of AI justified? Privacy and data security in the era of artificial intelligence". At Devōt, we are aware that the shift is happening, and getting data privacy and security right is the greatest barrier to AI's adoption. So, if you feel like you are unprepared for implementing AI or you are not sure what to explore, feel free to schedule a meeting with us, and let's bring AI to your project.
In the end, it's important to take humanity into account
Of course, adopting AI is a step towards building a more successful business, as it optimizes both operational efficiency and employee output. Using AI tools in workflows allows us to boost productivity by automating routine tasks and freeing up time. But let's not forget that artificial intelligence should primarily serve as support that makes our work easier and helps achieve a better balance between business and private life.
The goal of using AI is not to encourage toxic productivity, where the pursuit of more output can lead to burnout and dissatisfaction. Instead, it should aim to improve the quality of our work and life, enhancing our capabilities without replacing the human touch.
Organizations must understand the proper role of AI in the workplace. This involves not only integrating technology into our operations but also training leaders and teams to recognize the fine line between significant productivity gains and the risk of over-reliance. Ultimately, the successful implementation of AI will depend on our ability to balance these factors, ensuring that technology serves as a complement to human efforts, not a replacement.
