By: Devanathan Desikan, Engagement Partner - Product Engineering

The “What If…?” series explored what would happen in the Marvel multiverse if timelines of significant events differed from those in the original comics. It is fascinating to see how a slight change in an event could alter the entire storyline and the character’s perception.

Similar to Marvel’s multiverse, the software engineering universe is populated by various personas—product owner, scrum master, developer, architect, quality engineer, support engineer, and Site Reliability Engineer (SRE). Each has a defined role in delivering high-impact, high-quality software (event) to the market.

Imagine a parallel world in the software engineering life cycle where productivity is a competitive advantage, and enterprises continuously improve it. In this blog, we’ll talk about some of these possibilities and how businesses can achieve them.

Multiverse of possibilities

What if we nudge these engineering personas with key insights that change their day-to-day working methods to alter how they deliver software?What if developers could know the exact impact on code modules and build it right the first time rather than injecting unknown defects?

What if the scrum master could set up a pod with the required skills and provide contextual insights for better planning and estimation rather than deferring and re-planning?

What if quality engineers were not tail-end tuggers that ran an optimized test and influenced upstream quality?

What if the SRE could get a 360-degree view to design proactive resilience for faster issue resolution rather than relying on SMEs to triage prolonged issues?

While thinking of the possibilities is tempting, making them happen will be even more exciting.

Monetizing your Software Development Life Cycle (SDLC) data

Typical SDLC data is derived from multiple disparate sources. From the concept to the consumption of a business function, it resides in numerous integrated/semi-integrated tools. With increasing agility in development methodologies, the rate of change and the number of changes introduced have grown multifold. In fact, frequent iterations and DevOps cycles generate SDLC data with variety, volume, and immense value.

Data derived from these unstructured requirements and user stories, source code repositories, build and deployment logs, application logs, and test artifacts—scripts and scenarios, defects and incident data, release and change management tools, etc. have a wealth of information. SDLC data is one of the key outcomes of IT spending, but its true potential is not utilized for improving the outcome of software development.

The market continuously introduces new sophisticated products for automating SDLC and assisting software engineers. However, engineers are not provided the decision intelligence which will help them optimize and reduce rework and accelerate SLDC.

What if we start leveraging SDLC assets and monetizing their power to generate insights that could influence how we deliver software?

Injecting SDLC with insights to improve predictability and productivity

Vaccines provide insights into our internal organs and systems on potential threats so that the body can protect or prepare us to handle future invasions. Similarly, our software delivery life cycle needs to be injected with insights to ensure that we learn from the past or prepare for unfamiliar situations.

In the past few years, Artificial Intelligence (AI) and Natural Language Processing (NLP) have undergone a revolution with the transformer model and its manifestation in Large Language Models (LLMs). NLP algorithms can now decipher semantics and context of linguistic passages with the help of semantic text similarity, semantic search, and natural language generation. If this growing sophistication is leveraged in conjunction with SDLC, digital assets can derive these insights to make intelligent decisions faster and more accurately.

This will enable businesses to respond to all their SDLC queries. Each insight leveraged can enhance productivity through timely interventions, improve prediction quality, and help prioritize work, optimize efforts, and deliver projects faster.

Insights-led delivery can help business leaders in the following ways:

• Predictable results – Increasing business value throughput and revenue by delivering product/applications quicker than the current market standards.
• Improved quality – Enhancing trust in the outcomes delivered by the product/application can increase the number of end users onboarded, resulting in incremental revenue.
• Redirecting cost leverage – Reducing wastage by minimizing rework and accelerating insights creates opportunities to redirect cost to the right revenue-generating areas.

Leveraging AI in software delivery is still in its nascent stage. It is restricted to certain SDLC areas, predominantly in the automation and ITOps functions.

What If we extend the use of AI across the entire SDLC and change how we deliver software?

Improving your developers’ experience can benefit your customers

A frictionless developer experience is critical for improving product quality, delivery speed, and, above all, incubating creativity in the way of working. Insights embedded into the SDLC should focus on removing bottlenecks such as dependency on Subject Matter Expert (SME) to understand the change’s impact.

Too much information all at once, or too many simultaneous tasks results in cognitive overload. As a result, developers find it difficult to perform or process the information. It highly impacts the developers’ and software engineers’ general mental, emotional, and physical well-being. Many times, it leads to errors and lower quality of deliverables. Finding, reproducing, and fixing software issues is time-consuming, and increased mental switching cost can affect team’s morale and productivity.

What if developers had the proper insight to guide them to build first-time-right products, a kind of guardrail for development. What if they had answers to the following questions:

• Which are the impacted and dependent code modules?
• What functionalities are related to this code modules?
• What happened when this code module was last checked?
• Is it a defect-prone functionality?
• Is another pod member updating the same code?
• Are there any merge impacts?
• Whether there are hotspots in the code module for functional or resilience impacts?
• How we fixed this defect earlier?
• What to test and how much to test to access the regression impact?
• Could this code be built defect-free?

Providing a resilient application is the key to improving customer experience, and an insight-led software delivery model is the answer.

Knowing “What” to understand the impact of “What If…?”

Understanding what’s happening in the SDLC can be achieved by dashboarding factual data. The industry currently leverages multiple levers to accelerate this understanding, but here is what will help more.

• Know where to focus/invest – Understand the hotspots in your IT ecosystem to identify the right focus areas for increasing spend.
• Predictive delivery – Since business risk is more predictable, use this prediction to make product changes at the right time to capitalize on market conditions.
• Competitive advantage and strategic posturing – Leveraging AI-assisted productivity improvements to gain competitive advantage.

Using insight-led engineering can make endless possibilities come to life in the SDLC and possibly grant developers and businesses with mythical clairvoyance.

For more such insights please visit: Canvas by LTIMindtree.