✕ Thank You Thank you! A link to your copy of the report has been sent to the email address you provided. Read report
The gap between AI velocity and security
Nearly all development teams now rely on AI coding assistants, and the productivity gains are real. Yet most organizations are producing AI-generated code faster than they can review, secure, or govern it. The result: time saved in code creation is subsumed into greater efforts elsewhere in testing, review, and issue management phases, often negating benefits of AI-assisted development.
We partnered with third-party research firm UserEvidence to survey 831 software engineers and DevOps professionals. The findings reveal a clear inflection point: The question is no longer if you use AI coding assistants; it's how well you manage the AI-generated code they produce and establish deliberate, automated mechanisms to scale subsequent stages across the pipeline.
What you need to know
Of developers credit AI coding assistants with productivity and velocity gains
Saved per week on average by developers using AI coding assistants
Of respondents say additional security testing and remediation are the biggest bottlenecks
Increased likelihood of major efficiency gains with full governance in place
"The State of AI-Powered Software Development" highlights
See what our expert reveals about the report’s key insights and what they mean for your organization.
The complete AI-powered software development landscape
Productivity gains are real but are offset by friction
AI coding assistants are making it faster and easier to generate code at scale, but they create a bigger bottleneck to manage and secure that code.
Governance unlocks AI’s full potential
Structured AI governance lags behind developer demand for it, but those who have full governance in place are 55% more likely to see a major improvement in efficiency
Security concerns scale with AI usage
Higher utilization of AI-generated code yields both greater productivity and risk exposure.
AI-assisted development demands human oversight
Teams are ready to scale with AI-assisted security tooling, but they prefer to keep humans in the loop via pull requests or real-time IDE suggestions.
The teams that understand how to operationalize AI will win. To fully realize these gains, organizations must focus on guardrails and alignment to standards that serve as the blueprint for faster operations and remediation workflows.”
Keep pace with AI-powered software development pipelines
Black Duck Signal
Powered by ContextAI™, Black Duck Signal™ delivers real-time security analysis with human-curated intelligence so that security teams can govern, manage, and validate AI-assisted code without disrupting developer velocity.
Black Duck Polaris Platform
Black Duck Polaris™ Platform is a no-compromise AppSec platform unifying SAST, SCA, and DAST scanning with intelligent prioritization to deliver security at the speed, scale, and ambition of AI-powered development.
Black Duck Assist
Black Duck Assist™ delivers AI-driven remediation guidance backed by the industry’s most comprehensive open source knowledge base to provide accurate insights and real-time fixes within developer workflows.
FAQ
-
What is "The State of AI-Powered Software Development"?
It's a market research report produced by Black Duck in partnership with third-party research firm UserEvidence. In March 2026, we surveyed 831 software engineers and DevOps professionals to get a clear, data-driven picture of how AI coding assistants are reshaping development workflows—and where the real friction points lie. The result is a benchmarking resource that cuts through the hype and gives you the data your team needs to make informed decisions about AI adoption, governance, and security.
-
What does the report cover?
The report covers four interconnected themes that define where AI-powered development stands today. You'll find data on productivity—including the fact that AI coding assistants save developers an average of eight hours per week—alongside an honest look at the bottlenecks offsetting those gains: manual review, security testing, and code rework. It also digs into the governance gap, which is one of the most critical findings: only 30% of teams have full governance in place for AI coding assistant adoption, yet those who do are 55% more likely to see a major improvement in efficiency. Security risk rounds out the picture, with 64% of teams expressing moderate or extreme concern about AI-generated code introducing vulnerabilities.
-
Who should read this report?
If you're responsible for development velocity, application security, or engineering strategy at a medium- to large-size organization, this report is built for you. The survey respondents come from organizations with at least 500 employees—most with 2,000 or more—and more than half hold C-suite or senior leadership roles. Whether you're trying to make the case for AI governance investment, benchmark your team's adoption maturity, or understand the security implications of AI-generated code at scale, the data here speaks directly to your challenges.
-
Why does this research matter right now?
Because the adoption curve has already moved past the question of whether to use AI coding assistants—97% of survey respondents are actively using them today. The question your organization needs to answer is how to scale that usage without creating security debt or governance oversights you can't see until it's too late. AI is generating code faster than most teams can review, secure, or govern it. This report gives you the industry data to understand exactly where that gap exists and what the organizations closing it are doing differently.
-
What will I be able to do with the findings after I download it?
The report includes a deep profile of trending methods in AI-assisted development and concrete recommendations for building the governance and security infrastructure your AI-powered development program demands. You'll be able to benchmark your team's current AI adoption maturity against 831 of your peers, identify the specific workflow stages where AI-generated code is creating the most risk, and build a stronger internal business case for the tooling, processes, and oversight models that unlock AI's full efficiency potential—not just part of it.