Tricentis, a global leader in continuous testing and quality engineering, has released the findings of its global 2025 Quality Transformation Report.
The survey exposes the growing challenges in delivering quality software as well as dissonance among today’s technology leaders and professionals around how to prioritise speed, quality and cost in an AI-driven economy.
The research found that software quality falters amid competing pressures. As organisations chase productivity gains from Generative AI, teams are overwhelmingly focused on improving delivery speed (45%) over enhancing software quality (13%). As a result, 66% of global organisations are at significant risk of software outages in the next year, with 63% shipping code without fully testing it.
The pressure to shorten release cycles (46%) and accidental deployments of untested code (40%) are driving widespread quality compromises, with poor software quality increasing technical debt, maintenance costs, customer turnover, security breaches and compliance failures.
In the UK, the situation is worse, with 73% admitting to pushing untested code live – 12% more than US counterparts – and 44% reporting untested code accidentally slipping through. A vast majority of UK organisations (72%) still delay software releases due to a lack of confidence in test coverage, exposing persistent gaps in quality assurance.
The Quality Transformation Report is based on a survey of over 2,700 DevOps and quality assurance leaders and software developers worldwide. Respondents included CIOs, CTOs, and VPs of engineering from industries including the public sector, energy and utilities, manufacturing and financial services. The manufacturing and public sectors face the greatest outage exposure globally (69%), while in the UK, the energy and utility sectors are most at risk (69%), with financial services close behind (68%).
Other key findings include:
Quality gaps are costing organisations millions: Nearly half (42%) of global organisations believe poor software quality costs them US$1 million (£750,000) or more annually, with financial services firms reporting the steepest losses.
Misalignment between developers and leadership is blocking quality gains: Poor communication between software development and quality assurance (QA) teams (33%) and disconnect between leadership and software development teams (28%) are the top barriers to improving software quality globally. In the UK, however, pressure to release software too quickly (32%) is the biggest challenge, ahead of poor development and QA communication (30%) and leadership disconnect (26%).
Agentic AI looks set to help plug productivity, quality and performance gaps: The majority of organisations surveyed (82%) report excitement about the possibility of AI agents to assume their monotonous tasks in the development and delivery cycle, freeing up time for more strategic and rewarding work.
AI is gaining executive trust to drive critical high-stakes decisions: Nine in 10 CIOs, CTOs, and software delivery teams are confident in AI’s ability to autonomously make software release decisions.
Enterprises are gaining clarity on Generative AI’s business impact: Almost 90% of respondents say their organisations can effectively quantify the ROI from Generative AI within their software development lifecycles.
Strikingly, almost all (99.89%) respondents think autonomous testing will be useful for QA: Areas where technology leaders and software development professionals expect to see the most impact include improving software speed overall (28%), overall quality (28%), analysing test results (25%) and test case maintenance (23%).
“Recent software outages due to unchecked or untested code changes showcase just how critical high-quality software is to the wider organisational ecosystem. Having the right balance of quality and speed to serve developing technological needs is paramount,” said Kevin Thompson, Chief Executive Officer, Tricentis.
