Background

Requirements engineering is one of the most crucial stages in software design and development, as it addresses the critical problem of designing the right software for the customer.

For complex projects, collecting requirements for a product could take up to two years, as they typically come from a variety of disparate people and teams throughout the enterprise. This creates a large amount of variability and change when translating high-level business, marketing, or feature ideas into technical requirements and design details. This variability and continuous inflow of changes results in incorrect system architecture going unnoticed, causing inaccurate product information, longer time to market, costly re-work for the company, and poor customer experience.

According to Modus Operandi, the root cause of ~56% of requirements defects is the result of poorly written, unclear, ambiguous, or incorrect requirements. The remaining ~44% are due to incomplete and omitted requirements. And 82% of application rework tasks are related to requirements errors.

Technology Overview

Drona is an early-stage MVP developed by the University of Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT). It is designed and tested for (1) defining, modelling, and optimising technical requirements, (2) requirements authoring support, (3) reconciliation between requirements models, and (4) automated generation of full deliverables, as well as (5) acceptance criteria written as Gherkin test scenarios (syntax acclaimed by the software development community). The tool is built as an add-in for Enterprise Architect - a software system modelling platform used by more than a million companies worldwide. See Figure 1.

Drona was developed in collaboration with a large European financial actor, taking into account business analysts’ need for high-quality requirements.

With DRONA, customers are able to easily create flowchart models and benefit from AI-driven recommendations on how to make these models better.
This approach reconciles the models with textual requirements in real-time, and allows customers to provide textual requirements directly in the plug-in, and benefit from (1) requirements syntax check, (3) auto-completion, and (4) instant feedback about errors using NLP. The result is faster and more accurate requirements that are always in sync with the models that will end up being reviewed by business managers and support functions within the company.
The approach takes the requirements and models that the customer created with the help of Drona and automatically convert them into full deliverables that technical or any other partner teams can pick up and implement. The deliverables are tied one to one with the requirements, so there is no risk of the requirements being lost or misinterpreted in translation.
The approach automatically generate acceptance criteria. It takes the manual load off the product or feature owners’ shoulders, and it is automated. Thanks to the integrated structure, and that the requirements are known, they are then taken and used to generate both the acceptance criteria, as well as the test scenarios in an easy-to-understand language by the quality assurance engineers.

Benefits

Drona has five main benefits:

It provides an early understanding of design deficiencies, conflicts and changes during the requirements analysis and definition phases.

It ensures easy requirements capture and autocompletion generation when writing the requirements.

It provides a unified working environment that connects all stakeholders involved in the product or feature development.

It ensures a quick generation of the final deliverable without missing any information from the requirements text and models.

Opportunity

The university is seeking

testing partners for further validating the MVP, and/or
investment by an experienced seed capital fund

DRONA: Software Requirements Management Tool for Enterprise Architect

IP Status

Seeking