A new global survey from Leapwork underscores a growing tension in software development: while AI is widely viewed as essential to the future of testing, many teams remain hesitant to rely on it for mission-critical workflows.
Based on responses from more than 300 engineers and IT decision-makers, the research indicates that enthusiasm for AI-enabled testing is high. Nearly nine in ten respondents said AI now ranks as a priority within their testing strategy, and four out of five expect it to improve testing outcomes over the next two years.
Yet that optimism has not translated into broad, end-to-end deployment. Although 65% of those surveyed are experimenting with or using AI in at least some testing activities, only 12.6% have embedded AI across core test workflows. The data suggests that most organizations are proceeding deliberately rather than rushing into wholesale transformation.
Trust Issues
The trust factor is slowing adoption of AI. More than half of respondents said concerns about quality and dependability are limiting wider AI adoption. Testing teams reported that unstable or brittle tests, along with the difficulty of automating complex processes that span multiple systems, continue to create friction.
Updating test suites after changes to critical applications also remains time-consuming. 45% said it takes three days or longer to revise tests following significant system updates, delaying release cycles and dampening confidence in automation tools.
Manual effort remains a factor. On average, respondents said only 41% of their testing processes are currently automated. Test creation was cited as the largest obstacle to greater automation, followed by test maintenance. More than half pointed to lack of time as a primary barrier to advancing automation initiatives.
Larger Industry Concerns
Beyond testing, enterprise AI adoption appears to follow a similar trajectory. According to IDC, many companies are piloting AI initiatives, but far fewer have deployed them at scale in production settings. Factors including issues with guardrails and the need for greater IT expertise often slow the move from experimentation to full rollout.
In total, the data reflects a tech industry that is confident in AI’s long-term potential but still hesitant about its current limitations. Testing teams want faster cycles, broader coverage and reduced manual workload. But they are unwilling to sacrifice predictability to get there.
The survey also suggests that the path forward lies not in treating AI as a stand-alone solution, but in integrating it into resilient automation frameworks that teams already trust. As software systems grow more interconnected and updates become more frequent, reliability may prove to be the decisive factor that determines how quickly AI-driven testing scales from promising pilot to operational standard.
https://devops.com/survey-adoption-of-ai-software-testing-slowed-by-trust-issues/a>