This article draws on findings from a survey of 500 IT executives conducted by Capgemini and Sogeti in collaboration with Broadcom, as summarized by ZDNet.
The Pace of Modern Software Delivery
Continuous integration and continuous delivery (CI/CD) have fundamentally changed the rhythm of software releases. According to a survey of 500 IT executives released by Capgemini and Sogeti in collaboration with Broadcom, 58% of enterprises deploy a new build daily, and 26% do so at least hourly. When software ships that fast, quality assurance can no longer be an end-of-cycle event—it has to be woven into every stage of the pipeline.
That is the core argument for continuous testing: fast, automated validation of software releases that keeps pace with agile development. The survey found that nearly one-third of IT executives (32%) reported their IT departments had “fully embraced” continuous testing—a meaningful milestone, though it also means the majority have not.
Automation Gap: The Elephant in the Testing Room
It might seem self-evident that continuous testing demands heavy automation. The survey data, however, reveals a significant shortfall between aspiration and practice:
- Only 24% of test cases were being executed through automation.
- Only 24% of end-to-end business scenarios were covered by automated tests.
- Only 25% of required test data was being generated automatically.
Perhaps most telling: more than one-third of respondents (36%) said that the majority of their testing time is currently consumed by searching for, managing, maintaining, and generating test data. Test data management is frequently an afterthought in CI/CD strategy, yet here it emerges as one of the biggest drains on engineering time.
AI and Machine Learning: Potential, Not Yet Reality
Artificial intelligence is frequently cited as a solution to test optimization challenges, and the survey reflects this optimism—with caveats. At least 28% of executives reported an intent to introduce AI to manage their continuous testing processes. The survey authors note that machine learning capabilities could eventually allow systems to automatically determine which tests are required in each release and production cycle, reducing manual triage and prioritization effort.
However, the survey did not clearly quantify how deeply AI was already embedded in active testing cycles at the time of publication. The distinction between planned adoption and live deployment matters: intent is not implementation. Teams evaluating AI-assisted testing tools should look for concrete capabilities such as failure prediction, test case prioritization based on code change analysis, and self-healing locators for UI tests.
Transparency and Pipeline Visibility Are Non-Negotiable
Speed without visibility creates risk. The survey highlights that executives recognize this: 35% identified a “complete audit trail of testing activities” and a “consolidated test and release pipeline” as the most important test orchestration capabilities. A further 32% pointed to the need for a “single place for cross-team collaboration” and “continuous delivery pipeline visibility.”
These findings underscore that continuous testing is not only a technical challenge—it is an organizational one. When multiple teams contribute to a shared pipeline, siloed test results and opaque quality gates create bottlenecks and finger-pointing. A shared, real-time view of test status across the delivery pipeline reduces ambiguity and accelerates decision-making.
Capgemini’s Recommendations for Continuous Testing Success
Based on the survey findings, the Capgemini authors put forward four practical recommendations for organizations looking to mature their continuous testing practice:
- Create visibility of quality processes by implementing customized QA dashboards that surface test results and quality metrics across CI/CD pipelines in real time.
- Establish automated testing cycles early in the development lifecycle so that defects are caught as close to their point of introduction as possible—reducing the cost and effort of remediation.
- Leverage AI and machine learning to optimize test cycles, prioritize high-risk test cases, and reduce redundant test execution.
- Automate the self-provisioning of test data to eliminate the manual bottleneck that, per the survey, consumes over a third of testing time for many organizations.
What This Means for DevOps Teams
The survey paints a picture of an industry in transition. CI/CD adoption is widespread enough that daily and even hourly deployments are now common. Yet the testing infrastructure needed to validate that velocity safely has not kept pace. Automation coverage remains low, test data management is largely manual, and AI-assisted testing is still mostly aspirational.
For DevOps and platform engineering teams, the takeaway is clear: investing in test automation and test data infrastructure is not a quality-of-life improvement—it is a prerequisite for sustainable delivery speed. Without it, the pace of CI/CD becomes a liability rather than an advantage.
The original ZDNet article covering these survey findings can be found at zdnet.com.