Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
AI can help developers produce code faster without making a feature reach users sooner. Coding is only one part of delivery: changes still have to be reviewed, tested, integrated, and released. If AI increases the amount or size of work entering those stages, the team can end up with more code in progress rather than faster, more stable releases.
What “faster” means—and what it does not
Code generation speed, individual productivity, delivery throughput, and delivery stability are different measures. A developer may finish a task sooner, while the team’s rate of completed delivery stays flat or falls. And even when more changes reach production, that does not automatically mean releases are stable.
- Coding speed is how quickly code is produced.
- Individual productivity concerns a developer’s work and flow.
- Delivery throughput concerns how much software the team delivers over time.
- Delivery stability concerns the reliability of that delivery.
DORA’s 2024 overview reported that AI adoption significantly increased individual productivity, flow, and job satisfaction, while negatively affecting delivery throughput and stability. Its later 2025 summary reported positive relationships with throughput and product performance, but a continuing negative relationship with stability. These are report-level findings from different years, not a single before-and-after experiment or a guaranteed outcome for any team. Google Cloud’s 2025 report announcement says the study drew on responses from nearly 5,000 technology professionals worldwide and more than 100 hours of qualitative data; the available summaries do not provide enough methodological detail to compare every subgroup or reconcile the differing throughput findings.
Why faster code can create a slower delivery queue
When code is generated more quickly, review and validation do not necessarily speed up at the same rate. If developers submit larger changes or more work at once, reviewers must understand and assess more before a change can move forward. Testing, integration, and release controls also remain necessary.
#1 Best Overall
DORA’s generative AI summary, last updated April 13, 2026, says larger batches can take longer to review and are more prone to system instability. It also reports that a 25% increase in AI adoption was associated with a 1.5% decrease in delivery throughput and a 7.2% decrease in delivery stability. Those figures describe an association in DORA’s report summary, not a forecast for an individual company or proof that AI caused those changes. DORA’s summary attributes the relationship in part to increased change volume and larger batches.
The practical implication is not that AI inevitably slows shipping. It is that faster code production can shift pressure to the work around the code. Whether review, testing, integration, or release is the constraint depends on the team’s actual workflow; the report summaries do not establish one universal bottleneck.
What changed between DORA’s 2024 and 2025 findings
The two report summaries should not be collapsed into one timeless verdict. DORA’s 2024 overview linked AI adoption to higher individual productivity but lower delivery throughput and stability. Its 2025 report summary described positive relationships with throughput and product performance, alongside a negative relationship with stability. The reports therefore differ on throughput while continuing to flag stability as a concern.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →| Report | Reported relationship | How to read it |
|---|---|---|
| DORA 2024 overview | Higher individual productivity, flow, and job satisfaction; lower delivery throughput and stability. | Report-level findings; not proof of a causal effect for every team. DORA |
| DORA 2025 summary | Positive relationships with delivery throughput and product performance; negative relationship with delivery stability. | A later report with a different throughput finding, not a direct experimental measurement of what will happen in a particular organization. Google Cloud |
The 2025 report frames AI as an amplifier of existing organizational strengths and weaknesses. As DORA Lead Nathen Harvey and researcher Derek DeBellis put it in the report announcement: “AI doesn’t fix a team; it amplifies what’s already there.” The framing points attention to the delivery system around the tool, rather than treating generated-code volume as the goal. DORA’s 2025 overview likewise emphasizes the organizational system behind AI adoption.
Rank #3
How to find where your team is losing time
Measure the delivery path rather than judging success by how quickly code appears. Start with your team’s baseline, then track changes over time as AI use changes. DORA’s 2024 summary recommends baseline measurement and experimental continuous improvement; it does not establish that any one metric or intervention will work for every organization. DORA’s capabilities guidance discusses conditions such as testing, version control, fast feedback, loosely coupled architecture, and internal platform quality.
- Establish a baseline. Record how work moves from code change to release, including the time spent waiting as well as actively working. Use the same definitions and observation period when comparing later results.
- Inspect work in progress. Look for changes waiting for review, tests, integration, or release approval. Identify whether delays cluster at one stage instead of assuming that code generation is the limiting factor.
- Check batch size and review load. Compare how large changes are and how much review they require. If AI-assisted work is arriving in bigger batches, try smaller changes and see whether review and feedback become easier to manage.
- Examine feedback and test coverage. Automated tests and fast feedback can help teams detect problems earlier. Review whether tests run reliably and whether failures are addressed before changes accumulate.
- Review the surrounding system. Look at version control practices, architecture coupling, and the quality of internal platforms. These conditions can influence how easily changes are integrated and delivered.
- Change one practice at a time and measure again. Compare results with the baseline, including delivery stability. More code produced is not enough evidence that delivery improved.
What to prioritize when AI is part of the workflow
Keep the goal on reliably delivered changes, not maximum code output. DORA’s summaries support small batches, robust testing, version control, and fast feedback as useful conditions to examine. The 2025 overview’s broader point is that tools operate within an organizational system: a team with effective review and delivery practices may be better positioned to benefit than one whose existing queues and controls are already strained.
That is a reason to improve the workflow around AI, not a guarantee that a particular practice will solve a particular team’s delay. The report summaries establish associations and recommendations, but do not identify which stage is the largest bottleneck for every organization or prove the effect of an intervention at a specific company.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
