Guidelines and information on how to reduce our carbon footprint by changing our habits are easily available to everyone, and regulators have already moved to ensure industry makes its production practices more sustainable across the globe. As all sectors increasingly digitize, however, it has become clear that not only heavy industries such as construction, manufacturing and aviation are high Green House Gas (GHG) producers, but that the collective environmental footprint of energy preference, mobility choices and recycling policies by services businesses is also significant. In particular, as the world continues to become more digitalized, the environmental impact of software development and digital transformation can no longer be ignored.

In a bid to make software development more sustainable, more and more tech businesses are using “green coding”. Green coding is the practice that environmentally-aware software developers and integrators are applying across all their Software Development Life Cycle (SDLC) to help reduce the carbon footprint of their services, contributing to lowering their customers’ overall GHG emissions. As reporting standards become more rigid, with the EU Corporate Sustainability Reporting Directive (CSRD) and Corporate Sustainability Due Diligence Directive (CSDDD) for example looming ahead, providing accountability for emissions throughout the supply chain is an increasingly important proposition. These two regulatory requirements, although significantly reshaped by the recent Omnibus, define the onus of ESG reporting for larger companies, while also demanding greater transparency on supply chain emissions and double materiality. Add to these ISO 14001 standards or Green Software Foundation principles which promote software that reduces greenhouse gas emissions through greater energy efficiency, carbon awareness, hardware efficiency and more, and it’s clear the bar has been raised. Green coding is a simple concept; it is a practice that effectively uses less energy to produce, deploy and maintain applications, but how is this achieved in practice?

There are several best-practice KPIs to monitor sustainable coding: compute power and efficient CPU usage are some critical metrics that concur with high energy expenditure and rising carbon footprint when developing software or applications. Similarly, when developing an AI solution, it is important to reduce unnecessary throughput and ensure that paradigms use fewer cycles, as a huge amount of resources are mobilized with each invocation. Efficient prompt writing is one way to help not only reach the response to your query faster but reduce resource consumption and avoid memory overload.

Specifically, some key metrics to measure prompting efficiency are:

• Conciseness of prompts: concise prompts reduce token count and energy use, so knowing how much the prompt can be shortened without losing accuracy is critical.
• Latency: faster responses generally indicate fewer computational cycles.
• Compute Resource Utilization: CPU/GPU cycles and memory usage during inference.
• Success rate per iteration: the number of retries or reformulations needed to get the desired output. Inefficient prompts often require multiple attempts, increasing energy cost.

Overloaded memory is another typical energy consuming issue as garbage, including repeated and unnecessary global variables, weighs down the code. Another key optimization is writing clean, modular, and self-explanatory code to reduce maintenance complexity and avoid redundant computations, achieved by creating self-explanatory functions. These contain maintenance efforts to a specific area of code and enable faster debugging.

Fincons is committed to these clean coding best practices and weaves them into every step of the training process. From onboarding into the company and our Fincons Academies to mentorship initiatives, developers are tutored on these practices and encouraged to develop Design Thinking solutions to create more efficiencies and opportunities.

“As digital transformation continues to evolve with the increasing use of AI tools and solutions, we are focusing on reducing the environmental impact of our services to support businesses that want to curb emissions,” reports Riccardo Vescovi, Technology Innovation Hub Group Director at Fincons Group. “There’s a need to spread awareness on the fact that a correct architecture design and even coding it has very tangible, concrete repercussions on services, not only for the performance and costs but even balancing energy consumption,” continues Riccardo Vescovi.

“Machine Learning and Generative AI are useful for identifying best practices and to tailor on our specific needs, we recently conducted an internal hackathon involving our international engineering teams US & INDIA. This initiative helped us define our perspective and approaches to be adopted,” explains Nicola Lunanova, Technology Innovation Hub Manager and Senior Solution Architect at Fincons Group.

As sustainability criteria continue to become more important for businesses to respond to reporting requests and ever more critical consumers, choosing green coding partners can help cement and communicate these ESG efforts throughout the supply chain. “Companies that strive to be more environmentally friendly – concludes Riccardo Vescovi - should not dismiss the importance of green coding and need to take into serious consideration the impact of both new software implementation and of their legacy software and applications pool.”