The software industry is rapidly adopting artificial intelligence to write code, but it is facing significant challenges in ensuring the reliability of this code once it is deployed.
A survey involving 200 senior site-reliability and DevOps leaders from large enterprises in the U.S., U.K., and E.U. highlights the hidden costs of the AI coding surge. According to Lightrun’s 2026 State of AI-Powered Engineering Report, shared exclusively with VentureBeat, 43% of AI-generated code changes necessitate manual debugging in production environments even after clearing quality assurance and staging tests. None of the respondents stated that their organization could confirm an AI-suggested fix with just one redeploy cycle; 88% said they needed two to three cycles, while 11% required four to six.
The report’s findings come at a time when AI-generated code is expanding rapidly across global enterprises. Microsoft CEO Satya Nadella and Google CEO Sundar Pichai both claim that approximately a quarter of their companies’ code is now AI-generated. The AIOps market, which includes platforms and services for managing AI-driven operations, is valued at $18.95 billion in 2026 and is expected to reach $37.79 billion by 2031.
However, the report indicates that the infrastructure intended to catch AI-generated errors is lagging behind AI’s ability to produce them.
“The 0% figure signals that engineering is hitting a trust wall with AI adoption,” said Or Maimon, Lightrun’s chief business officer, referring to the survey’s finding that zero percent of engineering leaders described themselves as “very confident” that AI-generated code will function correctly once deployed. “While the industry’s emphasis on increased productivity has made AI a necessity, we are seeing a direct negative impact. As AI-generated code enters the system, it doesn’t just increase volume; it slows down the entire deployment pipeline.”
Amazon’s March outages showed what happens when AI-generated code ships without safeguards
The risks are no longer hypothetical. In early March 2026, Amazon experienced a series of notable outages that highlighted the kind of failure pattern described in the Lightrun survey. On March 2, Amazon.com faced a disruption lasting nearly six hours, leading to 120,000 lost orders and 1.6 million website errors. Three days later, on March 5, a more significant outage hit the storefront — lasting six hours and causing a 99% drop in U.S. order volume, with about 6.3 million lost orders. Both incidents were linked to AI-assisted code changes deployed to production without proper approval.
The aftermath was swift. Amazon initiated a 90-day code safety reset across 335 critical systems, and AI-assisted code changes now require approval from senior engineers before deployment.
Maimon highlighted the Amazon incidents. “This uncertainty isn’t based on a hypothesis,” he said. “We just need to look back to the start of March, when Amazon.com in North America went down due to an AI-assisted change being implemented without established safeguards.”
The Amazon cases demonstrate the key tension identified in the Lightrun report: AI tools can produce code at unprecedented speed, but the systems meant to validate, monitor, and trust that code in live environments have not kept up. Google’s 2025 DORA report supports this view, revealing that AI adoption is linked to increased code instability and that 30% of developers have little or no trust in AI-generated code.
Maimon referenced that research: “Google’s 2025 DORA report found that AI adoption correlates with an almost 10% increase in code instability. Our validation processes were built for the scale of human engineering, but today, engineers have become auditors for massive volumes of unfamiliar code.”
Developers are losing two days a week to debugging AI-generated code they didn’t write
The report’s most striking revelation is the amount of human capital devoted to AI-related verification tasks. Developers now spend an average of 38% of their work week — about two full days — on debugging, verification, and environment-specific troubleshooting, according to the survey. For 88% of the companies polled, this “reliability tax” consumes between 26% and 50% of their developers’ weekly capacity.
This is not the productivity gain that enterprise leaders anticipated when they invested in AI coding assistants. Instead, the engineering bottleneck has simply shifted. Code gets written faster, but it takes much longer to verify its functionality.
“In some senses, AI has made the debugging problem worse,” Maimon said. “The volume of change is overwhelming human validation, while the generated code itself frequently does not behave as expected when deployed in Production. AI coding agents cannot see how their code behaves in running environments.”
The redeployment issue adds to the time drain. Every surveyed organization requires multiple deployment cycles to verify a single AI-suggested fix — and according to Google’s 2025 DORA report, a single redeploy cycle takes a day to one week on average. In regulated industries such as healthcare and finance, deployment windows are often narrow, governed by mandated code freezes and strict change-management protocols. Requiring three or more cycles to validate a single AI fix can extend resolution timelines from days to weeks.
Maimon dismissed the notion that these multiple cycles represent prudent engineering discipline. “This is not discipline, but an expensive bottleneck and a symptom of the fact that AI-generated fixes are often unreliable,” he said. “If we can move from three cycles to one, we reclaim a massive portion of that 38% lost engineering capacity.”
AI monitoring tools can’t see what’s happening inside running applications — and that’s the real problem
While productivity loss is the most visible cost, the Lightrun report identifies a deeper structural issue: “the runtime visibility gap” — the inability of AI tools and existing monitoring systems to observe what is actually occurring inside running applications.
Sixty percent of the survey’s respondents cited a lack of visibility into live system behavior as the primary bottleneck in resolving production incidents. In 44% of cases where AI SRE or application performance monitoring tools attempted to investigate production issues, they failed due to the absence of necessary execution-level data — variable states, memory usage, request flow — that had not been captured initially.
The report depicts AI tools as operating essentially blind in the environments that matter most. Ninety-seven percent of engineering leaders said their AI SRE agents operate without significant visibility into what is actually happening in production. Approximately half of all companies (49%) reported their AI agents have only limited visibility into live execution states. Only 1% reported extensive visibility, and not a single respondent claimed full visibility.
This gap can turn a minor software bug into a costly outage. When an AI-suggested fix fails in production — as 43% of them do — engineers cannot depend on their AI tools to diagnose the problem, as those tools cannot observe the code’s real-time behavior. Instead, teams resort to what the report calls “tribal knowledge”: the institutional memory of senior engineers who have encountered similar problems before and can intuit the root cause from experience rather than data. The survey found that 54% of resolutions to high-severity incidents rely on tribal knowledge rather than diagnostic evidence from AI SREs or APMs.
In finance, 74% of engineering teams trust human intuition over AI diagnostics during serious incidents
The trust deficit is particularly pronounced in the finance sector. In a field where a single application error can lead to millions of dollars in losses per minute, the survey revealed that 74% of financial-services engineering teams rely on tribal knowledge over automated diagnostic data during serious incidents — significantly higher than the 44% figure in the technology sector.
“Finance is a heavily regulated, high-stakes environment where a single application error can cost millions of dollars per minute,” Maimon said. “The data shows that these teams simply do not trust AI not to make a dangerous mistake in their Production environments. This is a rational response to tool failure.”
The distrust extends beyond finance. A revealing data point from the report is that not a single organization surveyed — across any industry — has integrated its AI SRE tools into actual production workflows. Ninety percent remain in experimental or pilot mode. The remaining 10% evaluated AI SRE tools and opted not to adopt them at all. This highlights a significant gap between market enthusiasm and operational reality: enterprises are investing heavily in AI for IT operations, but the tools they acquire remain isolated from the environments where they would be most beneficial.
Maimon called this one of the report’s most significant findings. “Leaders are eager to adopt these new AI tools, but they don’t trust AI to touch live environments,” he said. “The lack of trust is shown in the data; 98% have lower trust in AI operating in production than in coding assistants.”
The observability industry built for human-speed engineering is falling short in the age of AI
The findings pose significant questions about the current generation of observability tools from major vendors like Datadog, Dynatrace, and Splunk. Seventy-seven percent of the engineering leaders surveyed expressed low or no confidence that their current observability stack provides sufficient information to support autonomous root cause analysis or automated incident remediation.
Maimon did not shy away from identifying the structural issue. “Major vendors often build ‘closed-garden’ ecosystems where their AI SREs can only reason over data collected by their own proprietary agents,” he said. “In a modern enterprise, teams typically have a multi-tool stack to provide full coverage. By forcing a team into a single-vendor silo, these tools create an uncomfortable dependency and a strategic liability: if the vendor’s data coverage is missing a specific layer, the AI is effectively blind to the root cause.”
The second issue, Maimon argued, is that current observability-backed AI SRE solutions offer only partial visibility — limited by what engineers chose to log at deployment time. Because failures rarely follow predefined paths, autonomous root cause analysis using only these tools will frequently miss the key diagnostic evidence. “To move toward true autonomous remediation,” he said, “the industry must shift toward AI SRE without vendor lock-in; AI SREs must be an active participant that can connect across the entire stack and interrogate live code to capture the ground truth of a failure as it happens.”
When asked what is needed to trust AI SREs, the survey’s respondents unanimously emphasized live runtime visibility. Fifty-eight percent said they require the ability to provide “evidence traces” of variables at the point of failure, and 42% cited the ability to verify a suggested fix before deployment. No respondents chose the ability to ingest multiple log sources or provide better natural language explanations — suggesting that engineering leaders do not want AI that communicates better, but AI that observes better.
The question is no longer whether to use AI for coding — it’s whether anyone can trust what it produces
The survey was conducted by Global Surveyz Research, an independent firm, and gathered responses from Directors, VPs, and C-level executives in SRE and DevOps roles at enterprises with 1,500 or more employees across the finance, technology, and information technology sectors. Responses were collected during January and February 2026, with questions randomized to prevent order bias.
Lightrun, backed by $110 million in funding from Accel and Insight Partners and serving clients such as AT&T, Citi, Microsoft, Salesforce, and UnitedHealth Group, has a vested interest in the problem the report outlines: the company offers a runtime observability platform designed to provide AI agents and human engineers with real-time visibility into live code execution. Its AI SRE product utilizes a Model Context Protocol connection to generate live diagnostic evidence at the point of failure without needing redeployment. This commercial interest does not detract from the survey’s findings, which are supported by independent research from Google DORA and the real-world evidence from the Amazon outages.
Collectively, they depict an industry facing a challenging paradox. AI has addressed the slowest part of software development — writing the code — only to reveal that writing was never the most challenging part. The real challenge was always knowing whether it works. And on that front, the engineers closest to the issue are not optimistic.
“If the live visibility gap is not closed, then teams are really just compounding instability through their adoption of AI,” Maimon said. “Organizations that don’t bridge this gap will find themselves stuck with long redeploy loops, to solve ever more complex challenges. They will lose their competitive speed to the very AI tools that were meant to provide it.”
The machines learned to write the code. Nobody taught them to watch it run.
