Balancing Build, Buy & Blend: Enterprise Application Transformation Strategies for the Age of AI -- OutSystems
By Lane F. Cooper, Editorial Director, BizTechReports
In today’s hyper-hybrid enterprise environments the traditional "build vs. buy" debate no longer makes sense. Instead, leaders must reexamine how well their organizations blend applications that are built internally, purchased and customized, and consumed as SaaS offerings in the cloud. While an overwhelming majority of organizations now use SaaS applications to run hundreds of enterprise applications across departments, it is also true that these functions must coexist with a broad array of critical software operating on private cloud, on-prem, and colocation resources.
So says Tiago Azevedo, Chief Information Officer of OutSystems, in a recent vidcast interview that explored how the emergence of low-code development platforms and the acceleration of generative AI are reshaping the way organizations modernize their IT architectures.
“It’s not about build or buy anymore — it’s all of the above,” he said. “And integration is now the core activity.”
Speaking after co-hosting a CIO Magazine executive roundtable on modernization strategies, Azevedo explained that CIOs today must navigate a “composable architecture” landscape — one that demands modularity, openness, and rapid integration across built, bought, and rented systems. Low-code platforms, he said, play a crucial role in stitching together these diverse elements.
Historically, businesses favored buying commercial off-the-shelf systems like SAP or Oracle, customizing them just enough to meet unique needs. That strategy worked when digital transformation simply meant automating manual processes. Today’s demands — fueled by relentless competition from nimble digital startups — require differentiation through in-house innovation.
“You can't buy your way into innovation,” Azevedo said. “You have to develop this capability yourself. Low-code platforms allow you to do that — fast, at scale, without sacrificing stability.”
By this, Azevedo means that companies cannot rely solely on purchasing commercial software to drive meaningful transformation. Prepackaged solutions are built to serve broad, common needs and rarely offer the differentiation required to compete in today's dynamic markets. True innovation, he argues, requires organizations to develop unique digital capabilities tailored to their specific strategies. Low-code platforms empower IT teams to build custom applications quickly and at enterprise scale, enabling businesses to innovate without the traditional trade-offs in time, cost, or stability that come with conventional development methods.
Putting Low-Code No-Code into Modern Context
He pointed to growing misconceptions about low-code technologies. Often labeled as tools for “citizen developers,” Azevedo stressed that today’s low-code environments are sophisticated enough for professional developers to build mission-critical, enterprise-grade applications. This evolution aligns with Gartner’s projection that by 2026, 80% of software development will involve low-code or no-code tools, up sharply from just 25% in 2020.
“We’re talking about real applications that support entire industries — not just workflows or simple forms.”
Azevedo also addressed the misconception that AI-generated code will soon replace traditional development altogether. While generative AI can write code based on natural language prompts — effectively speeding up development — it often lacks coherence in large, complex projects.
"You can generate content quickly, but you still need experts to structure, debug, and refine it,” he explained.
That is why the OutSystems platform, he said, has embedded AI deeply into its fabric through a module called Mentor — a generative AI engine that can build complete applications, including user interfaces, business logic, data models, and integrations. Mentor can even infuse intelligent agents into workflows, enabling, for example, automated summarization of client calls or dynamic email generation.
“Our customers are seeing results that seem almost too good to be true,” Azevedo said. He cited examples such as a major railway cargo operator that eschewed expensive SAP upgrades in favor of building lightweight, innovative applications using OutSystems. The company’s strategy led its CIO to win national recognition.
In another case, a major U.S. insurance firm sidestepped a 16-month, multi-million-dollar packaged software rollout by using OutSystems to deliver the same solution in just eight weeks.
Laying the Foundation for Application Generation Platforms
The rapid convergence of low-code/no-code development and generative AI is giving rise to a new category of technology known as application generation platforms. Traditionally, low-code and no-code tools have helped developers accelerate software creation by providing visual interfaces, reusable components, and built-in automation. However, the infusion of generative AI has pushed these capabilities even further. Spending on generative AI technologies is forecast to grow at a compound annual rate of 73.3% through 2027, reaching $143 billion globally, according to IDC — reflecting surging demand for AI-driven software solutions.
“Today, these platforms can take a simple prompt or business requirements and automatically generate a full application — including user interfaces, workflows, data models, and integrations. Developers then guide and refine the output, rather than building it manually.”
This leap in capability signals a shift in how organizations will approach software development. Rather than incrementally accelerating traditional coding processes, application generation platforms promise to fundamentally transform them — dramatically reducing the time, expertise, and resources needed to deliver enterprise-grade solutions.
“I expect industry analysts to recognize and track this new category as demand grows for tools that combine human expertise with AI-driven creation to meet the pressures of digital transformation at scale,” he said.
Human-centric Guardrail Enforcement
According to Azevedo, the secret to safely accelerating innovation at this pace lies in ensuring that responsible people are put in charge of setting and enforcement guardrails. Organizations must set policies around data access, application permissions, and development practices, automating governance and security throughout the software development lifecycle. McKinsey research supports this urgency, noting that companies that accelerate software delivery by even 20–30% can see revenue gains up to five times higher than their slower competitors.
“Low-code speeds up application development and enforces policies and compliance by design,” Azevedo said. “The productivity gains — sometimes 100 times faster than traditional methods — are transformational. It’s an amazing time to be in the market.”