Earlier this week, Apple announced at its annual Worldwide Developers Conference that it would rely in part on Google's Gemini models to power elements of its next-generation Siri experience.
After years of investing heavily in artificial intelligence, one of the world's most sophisticated technology companies decided that partnering made more sense than building everything itself. Why turn to an outside partner for such a visible part of Siri's evolution? Apple's reasoning isn't public knowledge, but the decision has highlighted a question that has been resurfacing across IT organizations:
As generative AI makes software development faster, cheaper and more accessible, should enterprises build more technology themselves? Or does the AI era actually strengthen the case for buying and integrating external capabilities?
CIOs have wrestled with build-versus-buy decisions for decades. Traditionally, the answer depended on cost, available talent, maintenance requirements and whether a capability offered genuine competitive differentiation. But now AI is changing many of those variables. What remains less clear is whether it changes the underlying decision itself.
Related:AI fuels a new wave of technical debt
For much of the modern enterprise era, the build-versus-buy debate was constrained by one simple reality: software development was expensive. Organizations needed specialized talent, lengthy development cycles and significant budgets to create and maintain custom applications. Purchasing commercial software was therefore often the easier, less risky option. Enter: generative AI.
Andreas Welsch, founder and chief human agentic AI officer at Intelligence Briefing, said AI helps to remove one of the biggest friction points facing IT organizations.
"For years, IT organizations have been struggling to keep up with requests for building new applications or improving existing ones," Welsch said. "The bottleneck was humans."
AI tools now accelerate the process of conceptualizing, building and maintaining applications, enabling teams to deliver more software than they could previously. The shift is already visible across enterprise software engineering. In its 2026 State of Code Developer Survey of more than 1,100 developers, code verification company Sonar found that 72% of developers who have tried AI coding tools now use them every day. It also reported that AI now accounts for 42% of committed code — a figure that is expected to reach 65% by 2027.
Those numbers help explain why the build-versus-buy conversation is being reopened. Tasks that once consumed weeks of developer time can increasingly be completed in hours; internal tools that previously failed a cost-benefit analysis may suddenly become viable projects.
Related:Why bank AI projects stall at approval
Nigel Duffy, CEO and founder of fintech company Cynch AI, said he believes AI is changing the economics of certain applications in particular. Historically, building and integrating internal alternatives was too costly. But today, he argued, some enterprises may find it easier to create niche applications tailored to specific business needs rather than add another third-party tool to an already crowded technology stack.
"[Generative AI] is best at building greenfield applications — and it is worst at integrating legacy third-party tools," Duffy said.
The temptation is to assume that because software is becoming easier to build, organizations should build more of it. But both experts cautioned that the equation is more complicated.
For years, IT organizations have been struggling to keep up with requests for building new applications or improving existing ones. The bottleneck was humans. — Andreas Welsch, founder and chief human agentic AI officer, Intelligence Briefing
The reality is that software development was never the only cost organizations faced. Creating an application may be getting easier but maintaining it remains difficult.
Related:The agentic shift at the Snowflake Summit: Finding a platform's 'right to win'
Many IT leaders have spent years trying to reduce technical debt, rationalize application portfolios and eliminate redundant systems. The prospect of dramatically increasing the number of internally developed applications may solve one problem while creating another.
"CIOs whose teams build applications in-house are assuming the risk," Welsch said.
Experienced CIOs continue to evaluate build-versus-buy decisions through a "total cost of ownership" lens, he added. While AI may reduce development effort, organizations still need to account for infrastructure expenses, cybersecurity requirements, testing, maintenance, support obligations, upgrades and ongoing enhancement work — and this quickly adds up.
The distinction matters because software complexity rarely disappears; rather, it moves. While AI may increase development output, it can also create new burdens for experienced engineers, who must now spend more time reviewing, validating and correcting AI-generated code. Productivity gains among less experienced developers might be offset by growing review and governance requirements among senior staff.
Duffy said he sees a related challenge emerging: staffing for this new era.
"The talent gap is moving more toward architecture and understanding of the business domain," he said. "This leads to a concentration of knowledge and expertise in a few key technical experts."
In other words, AI may reduce the scarcity of coding talent while increasing the value of people who understand systems, integrations, governance and business processes. Organizations that rush to build large numbers of internal applications may eventually discover that maintaining them requires expertise that is difficult to find.
That creates a different kind of dependency risk — but a dependency all the same. Rather than relying on external vendors, enterprises may become dependent on a handful of architects and domain experts who understand how internally developed systems actually work. In the long run, will this prove more restrictive?
Despite the attention focused on foundation models, relatively few organizations are likely to gain a meaningful competitive advantage from developing foundational AI technology themselves. The investment required to compete with companies such as OpenAI, Google, Anthropic and Meta is simply beyond the reach of most enterprises.
Instead, the strategic question increasingly revolves around how organizations apply AI to their own business context.
"Building foundation models is not a key competitive advantage for most enterprises, and they shouldn't imagine it is," Duffy said. "This is a moment to reflect on what your core competitive advantages are and lean into those."
That perspective challenges a common assumption emerging in many boardrooms: While AI makes software creation easier, that does not necessarily mean every layer of the technology stack deserves customization.
Welsch shared a similar view, recommending that CIOs focus on differentiation.
"We are still in the early phases of AI adoption," he said. "An organization's differentiation does not solely come from the foundational AI technology itself. It is rather the technology's application in a business context, in combination with an organization's data and semantics, that sets the organization apart."
The comparison to cloud computing may be instructive. Few enterprises today derive strategic advantage from owning data centers. Instead, competitive advantage comes from how organizations use their technology to improve customer experiences, streamline operations, or create new products and services. AI may ultimately follow a similar path.
For CIOs, the practical challenge is deciding where customization creates value and where standardization remains preferable.
Both experts point toward a similar framework. Commodity functions such as finance, HR, accounting and other highly standardized business processes remain strong candidates for commercial software. These applications benefit from mature support ecosystems, regulatory compliance capabilities and established maintenance models.
Areas tied directly to competitive differentiation may warrant a different approach. Organizations with unique workflows, proprietary data, specialized operational processes or distinctive customer experiences may increasingly find that AI makes custom development economically viable in ways that were previously difficult to justify.
Even then, the objective may not be building everything from scratch.
According to Duffy, enterprises will ultimately build significantly more software internally than they do today, particularly as AI lowers development costs. But he also warns that many organizations will underestimate the long-term complexity of owning those systems. In short: He believes they will make this choice, but not necessarily that it's a good choice.
Welsch is also skeptical of the long-term value of building too much internally: "To increase the efficiencies such as cost savings and time-to-value, CIOs should prioritize integrating and orchestrating AI capabilities rather than attempting to build them from scratch," he said.
That tension may ultimately define the next phase of enterprise technology strategy. AI is lowering the barriers to software creation, but it is not eliminating the tradeoffs that have shaped build-versus-buy decisions for decades.
For CIOs, the challenge has moved from determining whether they can build something themselves to whether they should — and whether the capability they're creating will still be worth owning years after the first version ships.
Senior Editor, InformationWeek
Madeleine Streets is a senior editor at InformationWeek, where she shapes stories and contributes news analysis through a CIO lens.
She comes to InformationWeek from TechTarget’s Learning Content team, in which she authored explainers and features on a range of enterprise IT topics. Before moving to the field of enterprise technology, Madeleine spent several years covering retail, consumer finance, and ecommerce technology for fashion trade publication Footwear News. She has also been published in Women’s Wear Daily, TIME, Associated Press, SELF, and Observer, among others. The thread that ties her coverage together is a commitment to honest, impactful storytelling -- and insatiable curiosity.
Outside of writing, Madeleine can be found studying wine, singing in her local choir, and working her way towards her annual reading goal of 100 books. She is based in New York City, US.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。