AI led Software Engineering Use Cases: Application to Requirements & Design

In the previous edition of the series, we have seen how AI transforms the software engineering lifecycle, specifically Management phases. In this edition we will see how Requirement engineering is affected by AI.

Requirements

The cost of errors in defining requirements increases exponentially as the project moves through life cycle stages hence the need to do requirements engineering in an effective and efficient manner. 

• Rational DOORS NG (Next Generation) is a leading requirement management tool from IBM. The IBM Engineering Requirements Quality Assistant module of the tool uses AI to leverage INCOSE (International Council on Systems Engineering) guidelines to evaluate, score and provide recommendations to improve requirements. It can help accelerate requirements review process and reduce training costs for junior requirements engineers. 

In financial services, non-compliances to some of the regulatory requirements can potentially result in fines of tens of billions of dollars to financial institutions. RegTech or Regulations Technology is a space where AI is being leveraged to ease the requirement gathering efforts and to help meet the regulatory requirements. A London based fintech CUBE helps financial enterprises identify global regulations and compliance requirements and automatically keep track of any changes to these regulations. Their platform can also help identify, in real-time, internal regulatory gaps in policies and procedures. Businesses engaging in cross-border selling might use their platform to gain an understanding of the rules and regulations for business across borders and build controls into internal company procedures. This can in turn reduce the requirement gathering efforts required for implementing these regulatory changes. Similarly a California based AI start-up, Ayasdi, can help financial enterprises understand anti-money laundering regulations requirements and adhere to them. 

• Source{d}, a spanish start-up, can perform analysis of organization codebase and help shape initiatives to rationalize and modernized the IT and achieve expected business velocity. It has mined 17 mil public repositories from 6.6 mil developers to understand natural language of code, intent of the developer and code similarities. Through its models it can devise useful insights about an organization’s coding assets, viz. extent of legacy code in the IT portfolio, code that isn‘t maintained, how often code is reused in the organization, who are the top-performing programmers, how effectively does the team collaborate, what key skills the organization lack in the teams and so on. 

Requirement prioritization, both functional and non-functional, is also seeing lot of research efforts. The methods being explored are related to Cumulative voting, Cost-Value method, Priority Grouping, Numerical assignment, Triage etc. They involve application of open algorithms such as Fuzzy Logic, Genetic Algorithm, case based ranking.

Design 

The IT industry has been trying to automate the software design and development activities for decades and AI probably holds the key to it. We are still nowhere near to accomplishing this goal but we have never been so near to it either. 

With the rising application of deep learning algorithms for processing images, the field of UX & UI, graphic design have seen a huge amount of progress in a short time. With the help of deep learning algorithms it is now possible for a designer to create clickable UI’s, compatible with popular mobile operating systems, directly from hand-drawn sketches on a piece of paper. This is called wire-framing automation and it is disrupting the product design workflow in a big way by reducing handover between UX designers and UI developers. Uizard, a Copenhagen based start-up, is doing just that to generate UI code targeting three different platforms, iOS, Android and web, from a single image input. Powering Uizard use case is the open sourced technology called pix2code available on github. To be sure, Uizard is not the first company to do that. Airbnb’s technology of wireframing automation enables it to create prototype code directly from whiteboard drawings. Sketch2code from Microsoft AI lab is another tool which provides similar feature.

• Designscape, a research project at University of Toronto, supported by Adobe and Microsoft, is about developing a system which aids the design process  by making interactive layout suggestions, i.e., changes in the position, scale, and alignment of elements and thereby helping a designer present different design options. 

In the area of graphic design, Adobe’s product suits like Adobe Sensei, Adobe Scene Stitch power intelligent features to help designers work effectively. The features include working with images, recognizing various elements inside images, automated tagging etc. There are plenty of companies, e.g. Grid, Wix, Squarespace which are leveraging AI to design the entire websites by taking some information from user as inputs. Tailor Brands and Turbologo uses AI to create logo without human intervention and by just taking answers to some questions as inputs. 

In the following articles we will see the AI led software engineering use cases in subsequent project lifecycles phases and also Sopra Steria’s offerings in this space.

Read more

AI led Software Engineering Use Cases: Application to Project Management activities