The adoption of AI in software development is continuously growing. According to the fresh data from Market.us Scoop, it is expected to reach $287 billion in ten years, with a compound annual growth rate of 21.5%. By the end of 2023, 45% of surveyed developers reported that they use generative AI in their workflows for measurable improvements such as a decrease in coding errors and cost savings. However, similar to any innovation, AI implementations in software development come with their risks. A Software Development and Engineering Manager and IEEE member Pratibha Sharma, currently working at Airbnb, shares her view on the AI role in software development and the issues companies face when trying to implement it.
Balancing Human Interventions and AI Applications
For instance, Pratibha Sharma notes that one of the main mistakes preventing companies from successfully implementing AI in their software development processes is their wrong perspective on the technology. “From the very beginning of the current AI proliferation wave, many companies still view it as the replacement of human developers, which establishes wrong expectations,” she explains. However, it is more productive to perceive AI as a tool that can take over routine work, freeing developers’ resources for more creative and strategic human-centered work.
This approach should be applied not only to the development process itself but to the final product as well if it involves AI applications in one form or another. During her tenure at Amazon, Pratibha Sharma was part of the team working on the customer service chatbot experience. One of the major factors of creating a product that answers the customers’ needs was determining, which elements of customer interactions could be easily automated, and which still need human intervention to be resolved. As a result, it became possible to process customer inquiries efficiently, saving human input only for unusual cases that cannot be processed automatically.
Nurturing the Teamwork
Another issue that leads to companies not unleashing the full potential of AI-based solutions in software development is the lack of integration. “It is not enough to provide developers with cutting-edge tools,” notes Pratibha Sharma. “They need to learn how to use them most productively, integrating them into their workflow.” Often it requires analyzing and reworking workflows, as well as ensuring that developers have the necessary training to use the new tools. In addition, organizations often require developing new metrics to evaluate their teams’ performance after they introduce new tools. For instance, more traditional metrics, such as lines of code or commits, become insufficient when generative AI is used to help with coding, and more goal-oriented criteria need to be established.
Implementing such an approach in practice requires productive interactions among teams with various specializations. While working at Amazon, Pratibha Sharma established partnerships with Product, Data Science, and Machine Learning Teams, which made it possible to create a productive environment for collaboration which was necessary for successfully releasing a final product. Pratibha Sharma adds that soft skills become of crucial importance for establishing productive teamwork around new technologies or tools. She mentions emotional intelligence, team development, and communication skills as those that helped her to increase her team’s productivity.
Combining Theory and Practice
It is also worth mentioning that to implement innovative technologies into their work processes successfully, one needs to work consciously, analyzing the potential impact of the changes. Pratibha Sharma follows this approach in her scientific publications, which are dedicated to the key aspects of the digital platform operation. She explores the risk management techniques in cloud infrastructures, as well as algorithms and strategies for fraud prevention that can be applied on online platforms, encompassing various solutions, including AI-based ones, and evaluating their effectiveness. These articles constitute an important contribution towards improving software development practices, as they highlight both theoretical and practical aspects of discussed topics, helping developers to find the best options.
“To succeed in such a rapidly changing domain as AI applications in software development one needs to learn constantly to keep up with the new technological developments,” concluded Pratibha Sharma. Throughout her career, she worked in several organizations, including Amazon, Lyft, and Airbnb, with each of them presenting its own task to solve within the realm of software development, which illustrates the versatility of her skills and her ability to bring value to any company she works at.
