AI for Requirements Management
(Source: NicoElNino /stock.adobe.com)
Artificial intelligence (AI) is officially on the march in the business world, with 33 percent of employees reporting that their organizations have started integrating new AI tools or technologies to improve business practices.1 Where these tools are applied often depends on where the biggest gains can be found.
To explore how AI can improve business practices, we spoke to Asif Sharif, the president and founder of Modern Requirements, a requirements life cycle solutions company. His company leverages their expertise in requirements management—the methodology development teams use to document, trace, analyze, prioritize, and establish requirements throughout product development life cycles.
Asif Sharif is an entrepreneur and visionary with decades of immersion in the information technology realm. He is the CEO and founder of Modern Requirements, a leading technology company revolutionizing requirements management with the award-winning tool, Modern Requirements4DevOps. In addition, his company recently developed and launched Copilot4DevOps, an AI solution for requirements management and DevOps. Asif is also the visionary behind Codeless ONE, a ground-breaking no-code platform designed for citizen developers. Asif's work in these different areas showcases his relentless dedication to transforming the technology landscape and empowering users at every level.
“In system development, requirements are the root of everything you do throughout the product’s life cycle,” explains Sharif. “The requirements define what is needed. And, if you get the requirements wrong, everything else goes wrong.”
Industry research confirms this assertion. According to the Project Management Institute, inaccurate requirements management is the most common reason that projects fail to meet their original goals. 2
In his 35 years in the field, Sharif has seen this unfortunate failure firsthand in a variety of industries, from medical devices to banking. He has also become all too familiar with the challenges that prevent companies from properly managing product requirements—and the technologies that can eliminate these problems.
Challenges to Requirements Management
When asked to name the most common challenge in requirements management, Sharif explains, “All the information you need to define a requirement is interconnected on multiple levels and dimensions, and this interconnectivity needs to be reflected in the system in order to manage requirements effectively.”
Another major challenge involves scaling requirements management tools and processes to fit the complexity of the product. “If you’re building an ultrasound machine or a financial system for an insurance company, those are complex systems,” says Sharif. “With so many more aspects and connections, defining requirements is much more difficult.”
The third challenge Sharif highlights is determining the length of a product’s development life cycle. “The length of the life cycle varies wildly depending on the complexity of the product,” he states. “If you’re a start-up, your life cycle view might be very short, and your model requirements might not extend beyond three months. A multiyear project, like an ultrasound machine, might take several years. So, your requirements management practices should not only adapt to the length of the life cycle but also pivot in response to any unforeseen circumstances that might arise during that time.”
In response to this challenge, Sharif recommends that teams preserve as much agility as possible while acknowledging the needs and constraints of the space in which they are working. “The intelligent approach is to understand what really works for you,” he says. “My advice is to be as light as you can and need to be—but no lighter.”
AI for Requirements Management
Most recently, Sharif applied the seemingly limitless power of AI to overcome some of the biggest challenges in requirements management, starting with the monumental task of writing system requirements.
Writing system requirements—whether for software, hardware, or firmware—takes a massive effort. “Many subject matter and technical experts have to converge to create the requirements. If the system is complex enough, this process may take months,” he explains. As Sharif’s team looked at this bottleneck, they began to ask how they could use AI to make this process more efficient.
His team’s solution involved applying large language models (LLMs) to the task using a tool Modern Requirements developed, called Copilot4DevOps, that was natively built as part of Azure DevOps.
“These LLMs are trained on nearly all publicly available human knowledge and are very smart,” Sharif explains. “This is the equivalent of having five Einsteins at your fingertips. So, we started to experiment with how to ensure that the system had the right context and information to intelligently, accurately define epics, features, user stories, test cases, and risk mitigations. And the results were stunning.”
According to Sharif, one chief technology officer from a client of Modern Requirements used the Copilot4DevOps extension to generate all the requirements for a new system they wanted to build—all in a single day.
“He turned to his product owners and said, ‘This is what I want us to build. Study this, fine-tune it, and let’s get going,” Sharif recalls. “In one day, he had accomplished what would have otherwise taken him and several experts months to do.”
This is only one example of how AI is accelerating the work of Sharif’s clients. Another is in the practice of determining the impact of a change to a project’s requirements.
As Sharif describes it, “Imagine that you want to make a change to a complex system, but first you want to know how this change could impact the system. Our impact assessment tool, having all the data and human input, scans the whole system and generates a detailed report on what areas will be impacted, and by how much, and how these impacts can be mitigated—all before you have time to get a cup of coffee.”
The Dynamic Future of Managing Requirements with AI
With such rapid gains in requirements management, thanks to AI and AI-powered technologies, Sharif is more than optimistic.
He believes AI is the future, and the efficiencies are unquestionable. In Sharif’s view, “Right now, people drive AI systems, but we are on the cusp of a model in which AI systems will drive AI systems. In this model, however, keeping humans in the loop will be critical. Human needs are always changing, and systems will need humans to engage with them in order to satisfy those changing needs. That will be the dynamic.”
“With this in mind, we need everyone to jump in and get their hands dirty with these technologies,” he continues. “Those who are quick learners, read and absorb new knowledge quickly, and work well with AI systems are the people who will be very successful in the future.”
1https://www.gallup.com/workplace/651203/workplace-answering-big-questions.aspx#
Asif Sharif president and founder of Modern Requirements.