Software testing stories
More than half of engineering teams are now using AI to write code, but weak oversight is leaving security, dependency and performance risks in production.
The scheme has kept more than 30 neurodivergent people in technology roles, as DXC expands support and AI training for the next UK cohort.
Developers can now avoid manual test setup as Kong's new link keeps API definitions, environments and credentials aligned across Insomnia and Konnect.
AI-generated code is outpacing enterprise review processes, prompting Qodo to add tools that flag cross-repo risks and enforce standards.
By focusing on evidence and small reversible changes, loop engineering could curb costly AI coding mistakes before they reach production.
Strict controls are now central to One NZ's AI push as it guards customer data and avoids costly errors in billing and finance.
The survey suggests employers now fear junior coders can generate output with AI, yet still struggle to explain or debug their own code.
Weaker oversight could turn AI-generated code into a costly drag, with security flaws and technical debt rising in enterprise projects.
AI pilots are faltering where firms still judge success by hours saved, leaving customer value and workforce design unresolved.
Enterprises modernising software delivery could cut testing risk and speed releases as the firms pair consulting with AI-enabled quality tools.
The pilot is intended to help firms prove AI is being managed safely and consistently as they move from trials to large-scale use.
Firms risk costly missteps as automated hiring filters miss staff who could be retrained for AI-augmented roles.
Yet live deployments are causing headaches for engineering teams, with most respondents reporting more incidents and heavier rework after AI code goes live.
Enterprises trying to cut software maintenance costs may find more automation, as Tech Mahindra repositions legacy application work around agentic AI.
The new planning tool aims to cut bugs and security flaws before code is written, as the startup's seed funding reaches USD $17 million.
False negatives from automated scanning tools are fuelling a shift towards human-led AI security testing across large organisations.
Software teams could catch regressions before release as the new verifier checks pull requests against live production behaviour inside existing workflow tools.
UK businesses struggling to push AI pilots into production will get onshore support from a merged consultancy focused on delivery, quality and security.
Enterprise software teams are far more willing to use AI before production, with trust dropping from 82% at build to 58% at release.
The restricted model could speed up vulnerability fixes across Cohesity's platform as AI intensifies both attack and defence in critical software.