EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
Tech Trend: enhancing software development with GenAI
In the second episode of our ‘Tech Trends 2024’ series, we dive deep into the transformative potential of Artificial Intelligence (AI) augmented software development. Join Radhika Saigal, Partner, Technology Consulting, EY India, as she helps uncover the profound impact of AI on software engineering excellence, productivity gains, resulting sustainability challenges, ethical considerations, and the path ahead towards successful GenAI adoption.
For your convenience, a full text transcript of this podcast is available on the link below:
Welcome back to EY India Insights podcast. I am Pallavi and we are happy to be back with another episode of our latest series ‘Tech Trends 2024’. Throughout the series, we aim to delve deep into the upcoming tech landscape along with our leading tech consulting partners. In today's episode, we will be exploring the transformative impact of Artificial Intelligence (AI) augmented software development, ushering us into a new era of efficiency and innovation.
Stay tuned as we uncover how GenAI tools such as GitHub’s Copilot and Jasper are reshaping the landscape, boosting developer productivity, and automating various processes in DevOps. Joining us today is Radhika Sehgal, Partner, Technology Consulting, EY India, bringing over two decades of experience in spearheading complex digital transformation across industries. Radhika, thank you for joining us today and welcome to our podcast.
Radhika: Thank you, Pallavi for having me here.
Pallavi: There is a lot of discussion on generative AI (GenAI). According to you, why is that?
Radhika: AI is not new. The term artificial intelligence was introduced by John McCarthy in the year 1956, before many of us were born. In the year 1966, first chatbot Eliza was launched. Fast forward to 1997, IBM’s Deep Blue defeated chess champion Garry Kasparov, which caused a big revolution. Siri got introduced in the year 2011 on the Apple iPhone, and Alexa entered our household in 2014. GPT-3 was introduced in 2020, followed by ChatGPT becoming public in 2022. So, AI was here for more than 75 years, but the pace of adoption is very, very fast.
When we look at Twitter, which took two years to get to 1 million users, ChatGPT crossed 1 million users in just five days. AI, machine learning, deep learning, and now it is generative AI where the deep learning is used to generate text, audio, and video. Now, this is disrupting innovation, so the talk continues.
Pallavi: Thank you, Radhika. How does AI contribute to software engineering excellence?
Radhika: AI is helping software development front to back. From GenAI helping in auto code generation, bug detection, testing as well as bug resolution. AI-powered code generations are also being used for pair programing as we speak. AI is being leveraged for predictive analytics to predict how certain changes in the code will impact performance, as well as user acceptance.
AI can use and determine how the user behavior is changing and give insights on how hyper personalized user interface (UI) can be rendered. It is also being integrated with machine learning models (MLMs) and helps improving performance over time for the various software products.
AI is also improving the full DevSecOps process, from building to debug to merger, to getting the auto deploy, as well as automating the entire Continuous Integration/Continuous Deployment (CI/CD) pipeline.
AI is helping in detecting and mitigating vulnerabilities, anomaly detections; it is helping in proactive software monitoring, root cause analysis, as well as self-healing. Beyond software development and DevSecOps, it is also being used in terms of business requirement gathering.
Business case generation from historical cases is now common. The feasibility analysis, the auto generation of user codes as well as the project plans - in all of these, we are leveraging GenAI. Prototype generations, converting user stories to data backlog, these are also now enabled using GenAI. So, we have the entire lifecycle of the coding being done, DevSecOps being done with the help of GenAI. We are also seeing a lot of use cases of GenAI in cloud FinOps, and we are helping a few of our clients implement the same.
Pallavi: What is the productivity gain that we can anticipate with this AI software engineering excellence?
Radhika: Overall, productivity gain is in the range of 25-30% on an average. The improvement in business gathering could be around 20%, coding and testing maybe around 60%, and for deployment around 25 to 30%, resulting in overall Software Development Life Cycle (SDLC)- level productivity gain of 25-30%. It also depends upon the type of software being developed. We see improvement is much higher in mundane manual tasks or repetitive tasks, compared to those that require innovative solutions.
Pallavi: Thank you, Radhika. Turning our attention to the sustainability challenges faced by the software engineering industry, what are the key challenges that you see in this area?
Radhika: Sustainability is an issue, for sure. As per MIT research, training a single AI model can emit as much carbon footprint as five cars in their lifetime. So, looking at AI from a sustainability lens, from an ESG standpoint, it is extremely crucial.
Pallavi: As you outline the challenges, could you also tell us how one can leverage GenAI in a more sustainable fashion?
Radhika: Firms are now being cognizant about the fact that GenAI is contributing to carbon footprint, and there are sustainability challenges. Cloud service providers are providing serverless technology to eliminate the resources, data management tools, as well as data sets in order to reduce the carbon footprint.
As software engineers, we can do a lot to ensure that we are leveraging GenAI in a sustainable fashion. First, use existing trained models rather than training a new model. Leverage composite solutions of trained models to solve complex problems. Third, Large Language Models (LLMs) are not always the right solution for the problem, so consider the best approach, including leveraging lighter weight models to solve the business problem.
Last, but not the least, GenAI is automating and optimizing coding in a way that increases efficiency, and that itself is reducing the carbon footprint of manual coding. We, as software developers, need to be cognizant and should adopt these methodologies to ensure that we are creating a sustainable future.
Pallavi: As we pivot towards GenAI, ethical considerations are paramount in AI-driven software engineering practices. How do we ensure these considerations are integrated effectively?
Radhika: AI is as efficient as our data. If we have past data that has biases, it can introduce bias as well as ethical concerns in the listening. It is extremely important that we have clearly documented ethical guidelines covering privacy, security, transparency, and bias. This needs to be there in every business, in every organization, well documented and understood.
Second, we have to have a clear mechanism for transparency as well as traceability to ensure that we have explainability for every decision making that is being done through AI modeling. This is required both for internal audit as well as for regulatory purposes.
Last, but not least, continuous monitoring and proactive mechanism to detect, rectify any ethical concerns are paramount.
Pallavi: What are the other challenges that we face in the adoption of GenAI and are there any solutions for the same?
Radhika: One of the biggest pain points would always remain data quality and hence that needs to be addressed primarily – from having the right data, using tools to having robust data governance, as well as a domain expertise to understand any issues in terms of both data and output – are very important.
Having the right AI strategy is the second one. The tech-forward strategy, ensuring that we have modern API architecture, less tech - all of those engineering and principles around modern tech architecture remain central. We should have AI embedded in it in a way that we are facilitating best technology strategy embedded with AI.
Regulatory requirements have to be taken into consideration; data reliability, compliance with regulatory directives; we also spoke about biases and fairness - all of these are important elements that we need to take into consideration.
Some of the businesses are not getting leadership buy in. Ensuring that we buy in top-down right from getting AI to get embedded into an existing ecosystem; it is very important that we start looking at it from a pilot project perspective, highlighting the competitive landscape.
While we do know that data security remains an issue irrespective of whether you are using a GenAI or not, but when you are using an AI or GenAI tool, it increases the usage of data and reliance on data. So Personally Identifiable Information (PII), data, data tokenization, data masking, data-level access control, all of these remain important when you use AI.
Last, but not the least - skills shortage. This is a new technology and will continue to change. GenZ will have newer technology coming in, but I see this as an opportunity rather than a challenge. This gives every engineer right now to immediately adopt to this new technology, innovate themselves and disrupt themselves rather than being disrupted.
So, these are some of the principles that I would advise any software engineer to pay attention to.
Pallavi: Thank you, Radhika for those valuable insights. I am sure all our listeners would have got a clear idea of how AI can augment software development. It has been an enlightening conversation.
Radhika: Thank you, Pallavi.
Pallavi: On that note, we come to the end of this episode. Thank you to all our listeners for joining us on this EY India insights podcast Tech Trends 2024 series. Stay tuned for more episodes. Until next time, this is Pallavi, signing off.
If you would like to listen to our podcasts on the go:
