Virtual assistants in QA (VA) are generally described as smart agents, which use data and statistical methods to evaluate current and historical activities of software process workers, to recommend next best actions.
According to Iosif Itkin, et. al., of Exactpro Systems (Exactpro), the ability of a human operator to analyse and make decisions based on data driven testing is becoming a bottleneck.
In June 2019, QA Vector® Research asserted that AI-enabled software risk management remained aspirational. At that time, fewer than 5% of FinTech and eCommerce firms had embedded virtual assistants into their software quality assurance practices.
If the goal of software quality assurance is to find the “hidden wreckage and pollutants”, i.e. technical debt, that have a material impact on overall software performance, then VA may accelerate time-to-value.
Exactpro’s proposal for VA may well ease the burden on quality leaders for determining where to prioritise QA time and effort, thereby enabling the pursuit of mission critical assurance.
To develop an effective VA strategy, Exactpro suggest that quality leaders should focus on:
- fuzzy logic, such as natural language processing, to understand code base inputs;
- context sensing, including domain knowledge, to create system knowledge; and
- log generation, containing states and semantics, to generate meaningful defect reports.
Download this vision paper on the development of an intelligent virtual assistant for software testing teams to discover how to leverage assistive technology across the software development life cycle.
In addition, Exactpro recommend that intelligent software quality agents should interface directly with AI enabled defect reporting systems, for an enterprise view of software quality.
Quality leaders at financial firms should also review delivering quality for financial infrastructure firms, to reinvent the quality dimension and drive business engagement.
