DORA 2024 found a direct correlation between internal developer platform quality and the ability to extract value from AI tools. That finding reframes platform investment as AI readiness.
There is a finding buried in the DORA 2024 report that I think most AI adoption coverage has missed. The report found that 90 percent of organizations have adopted at least one platform engineering practice. That surface number sounds encouraging. What follows it is the part that matters: there is a direct correlation between the quality of an organization's internal developer platform and its ability to extract value from AI development tools. Not a loose association. A direct correlation. Organizations with high-quality platforms are getting value from AI. Organizations with poor platform quality see AI as a burden rather than an enabler. Platform quality is no longer just a developer experience concern. It has become a precondition for AI value capture.
I want to be precise about what an internal developer platform is, because the term gets used loosely enough that it sometimes loses meaning. At its core, an internal developer platform is the organizational infrastructure that sits between the raw cloud and the engineer writing code. It typically includes what practitioners call "golden paths": opinionated, pre-configured project templates and deployment pipelines that let a team start building a new service without also needing to become an expert in container orchestration, secret management, and access control. It includes self-service infrastructure provisioning, where a developer can spin up a test database or a new environment through a catalog interface without filing a ticket with a central team. It includes an integrated toolchain where version control, CI/CD pipelines, test execution, observability dashboards, and access controls are connected and configured to work together. The goal is to reduce the cognitive overhead of getting from working code to running service, and to make the responsible path also the easy path.
The DORA 2024 connection to AI value is not incidental. The report found that AI adoption improves deployment throughput but hurts delivery stability, and that the mechanism is exposure: AI accelerates code generation and exposes weaknesses in the surrounding infrastructure when that infrastructure cannot catch mistakes automatically. A mature internal developer platform is exactly what catches those mistakes. When a developer uses an AI code generation tool on a team with high-quality platform infrastructure, the generated code flows through automated tests before it can merge, it deploys through a pipeline that runs integration checks, and it is monitored by observability tooling that surfaces anomalies in production within minutes. The platform is the filter that makes AI acceleration safe. Without the filter, you get the acceleration and the instability together.
Barney's (1991) resource-based view of the firm is the IS theory that I find most useful here. Barney argued that sustainable competitive advantage comes from resources that are valuable, rare, difficult to imitate, and non-substitutable. A high-quality internal developer platform meets all four criteria. It is valuable because it directly affects delivery speed, reliability, and now AI value capture. It is rare because building one requires years of accumulated engineering investment and organizational discipline that most organizations have not made. It is difficult to imitate because the platform is not a product you can purchase from a vendor. It is an organizational capability built from decisions made over time: the decision to invest in test automation when it was slow, to build self-service infrastructure when ticketing was easier, to standardize tooling when diversity felt more flexible. A competitor can buy the same developer tools. They cannot quickly buy the platform quality that comes from a decade of consistent engineering investment. It is non-substitutable because no amount of AI tooling compensates for the absence of the infrastructure that makes AI tooling safe.
The McKinsey State of AI 2025 report found that 88 percent of organizations say they use AI, but only 7 percent report having fully scaled AI across the enterprise. The gap between those numbers is large enough to carry a lot of explanatory weight, and I think a meaningful share of that gap is platform quality. Organizations with weak internal developer platforms are running AI tools as individual experiments, not organizational capabilities. One team uses one AI coding assistant. Another team uses a different one. The two codebases have different testing standards, deployment practices, and observability setups. There is no consistent infrastructure that makes AI assistance reliable and safe across the organization. This is the difference between AI as something that some individuals happen to use and AI as something the organization can depend on. The first category is easy to count in a survey. The second category requires platform quality.
Trist and Bamforth's (1951) sociotechnical systems framework argued that technical systems and social systems are interdependent, and that optimizing one at the expense of the other produces suboptimal outcomes for both. In the context of internal developer platforms and AI tools, the technical system is the platform infrastructure: the pipelines, the tests, the observability, the golden paths. The social system is the team: the norms around code review, the culture of treating test coverage as a shared responsibility, the shared understanding of what "ready to deploy" means. Both have to be healthy. A technically excellent platform deployed into a team that treats testing as optional and code review as a formality will not catch AI-generated mistakes effectively, because the humans are not using the technical controls the platform provides. A technically weak platform in a team with strong engineering discipline will produce better outcomes than a technically weak platform with weak discipline, but the ceiling will still be low. The DORA finding that platform quality correlates with AI value is measuring both dimensions at once, because platform quality in DORA's framework includes the organizational practices that determine whether the technical controls are actually used.
Gartner's April 2026 forecast projects AI agent software growing from $86.4 billion in 2025 to $206.5 billion in 2026. The organizations that will extract value from that spending are not the ones with the biggest AI budgets. They are the ones that invested in platform quality before the AI tooling arrived, and are now in a position to absorb AI capabilities into an infrastructure that makes them safe. The organizations that are buying AI tools to fix a platform problem they have not yet acknowledged are going to generate exactly the kind of data the DORA 2024 burden finding describes. The tools arrive. The instability follows. The platform investment gets proposed as a solution to the problem the AI tools made visible. The right sequence is to build the platform, then amplify it with AI. Almost no organization does it that way. I think that is what IS research needs to study more carefully: not just whether organizations adopt AI, but whether the infrastructure conditions for safe AI adoption were in place before adoption happened.
---
claims_checked:
- "DORA 2024: 90% of organizations adopted at least one platform engineering practice": "https://dora.dev/research/2024/dora-report/"
- "DORA 2024: Direct correlation between platform quality and AI value realization": "https://dora.dev/research/2024/dora-report/"
- "DORA 2024: Organizations with poor platform quality see AI as a burden, not an enabler": "https://dora.dev/research/2024/dora-report/"
- "DORA 2024: AI adoption improves throughput but hurts delivery stability": "https://dora.dev/research/2024/dora-report/"
- "McKinsey State of AI 2025: 88% use AI, 7% fully scaled": "https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai"
- "Gartner: AI agent software $86.4B (2025), $206.5B (2026)": "https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-forecasts-worldwide-it-spending-to-grow-9-8-percent-in-2026"
claims_unverified:
- "Barney (1991) resource-based view: drawn from IS theory background, not re-fetched this session"
- "Trist and Bamforth (1951) sociotechnical systems: drawn from IS theory background, not re-fetched this session"
sources_used:
- "https://dora.dev/research/2024/dora-report/"
- "https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai"
- "https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-forecasts-worldwide-it-spending-to-grow-9-8-percent-in-2026"
word_count: 1080
About the author
Share
More notes
Related notes