For decades, the fundamental question of whether to "build" custom software solutions or "buy" off-the-shelf products has been a cornerstone of strategic decision-making within the technology industry. This enduring dilemma transcends the boundaries of traditional IT departments, permeating every modern product team. While the build-versus-buy calculus has always been multifaceted, factoring in development costs, time-to-market, functionality limitations, and maintenance overheads, the advent of generative artificial intelligence (AI) and advanced user programming tools is now introducing unprecedented complexities and opportunities, reshaping the landscape of enterprise software solutions.
The Enduring Dilemma: Build vs. Buy
The core of the build-versus-buy debate often hinges on a company’s strategic priorities and core competencies. Historically, if a software solution addresses a problem central to a company’s competitive advantage or defines its unique operational workflow, the inclination has been to build. This approach allows for granular control, tailored functionality, and seamless integration with existing proprietary systems. However, building custom solutions is notoriously resource-intensive, demanding significant upfront investment in development, extensive testing, and continuous maintenance and updates. The risks associated with project delays, budget overruns, and the eventual accumulation of technical debt are substantial.
Conversely, opting to buy a commercial off-the-shelf (COTS) solution, particularly a Software-as-a-Service (SaaS) offering, typically promises faster deployment, lower initial costs, and access to a vendor’s specialized expertise and ongoing support. SaaS adoption has skyrocketed over the past decade, with market reports indicating a compound annual growth rate (CAGR) of over 15% and projections estimating the global SaaS market to reach nearly $200 billion by 2024. Companies leverage SaaS for functions ranging from customer relationship management (CRM) and enterprise resource planning (ERP) to human resources (HR) and cybersecurity. However, purchased solutions often come with inherent limitations in customization, potentially forcing businesses to adapt their processes to the software rather than the other way around. This can lead to vendor lock-in, difficulties in integrating with unique legacy systems, and ongoing subscription costs that can escalate over time.
For many large enterprises, the reality has long been a hybrid approach, where "buy" often necessitates significant "customize" efforts. Enterprise-grade SaaS platforms, while powerful, frequently require extensive configuration, integration, and even custom development to align with the specific operational nuances, compliance requirements, and data architectures of a particular business. This often blurs the lines between a pure buy and a pure build strategy, creating a complex interplay of vendor-provided functionalities and in-house adaptations.
A Historical Trajectory of User Empowerment: The Rise of User Programming
The ability for non-technical users to create software solutions, often referred to as "user programming," has a surprisingly long and impactful history, predating the current AI revolution by decades. This movement was born out of a desire to alleviate the bottlenecks inherent in traditional IT departments, which were often overwhelmed by a deluge of requests from "the business" for custom applications, reports, and data manipulations.
The paradigm shift truly began in 1979 with the invention of VisiCalc, the first spreadsheet program designed for personal computers, specifically the Apple II. VisiCalc empowered accountants, financial analysts, and business managers to create complex financial models and calculations without needing to write traditional code. This was a monumental leap, democratizing a form of programming and giving non-technical professionals direct control over their data and analytical tools. The impact was profound; VisiCalc is often credited with catalyzing the personal computer revolution, transforming PCs from hobbyist gadgets into indispensable business instruments.
The legacy of VisiCalc continued through the widespread adoption of Microsoft Excel, which became the de facto standard for data manipulation across virtually every industry. Millions of user-created programs, predominantly in the form of complex formulas and macros, run daily within Excel. This era further evolved with the introduction of Visual Basic (VB) in 1991, which provided a more structured yet still accessible environment for building graphical user interface (GUI) applications. Visual Basic for Applications (VBA) embedded within Microsoft Office applications extended this capability, allowing users to automate tasks and build custom functionalities directly within their familiar productivity suites. These tools, particularly VBA, can be seen as early progenitors of the low-code movement, enabling millions of individuals to develop custom solutions with minimal traditional programming expertise.
The past decade has seen a significant acceleration in the low-code and no-code (LCNC) movement. Platforms like OutSystems, Mendix, Appian, and Salesforce’s Lightning Platform have emerged, offering visual development environments that allow business users and citizen developers to build sophisticated applications through drag-and-drop interfaces, pre-built components, and intuitive workflows. This wave aimed to bridge the gap between business needs and IT capacity, further empowering non-technical personnel to innovate and solve problems more rapidly. Industry reports from Gartner predict that low-code development will account for over 65% of application development activity by 2024, highlighting its growing prominence.
The Generative AI Revolution and Natural Language Programming
While low-code and no-code tools significantly lowered the barrier to entry for application development, they still required a degree of logical thinking, understanding of data models, and familiarity with platform-specific interfaces. The advent of generative AI, however, represents a new frontier in user programming, potentially making solution creation accessible to an even broader audience.
Generative AI models, exemplified by large language models (LLMs) like OpenAI’s GPT series or Anthropic’s Claude, are fundamentally changing the interface for creating software. Products like Lovable and Bolt, as mentioned, are pioneering a new generation of user programming tools where the "programming language" is essentially natural human language—English, or any other spoken language. This capability, sometimes termed "vibe coding" or "prompt engineering," allows users to describe their desired application or workflow in plain text, and the AI system generates the underlying code, components, or entire application structures.
The implications are transformative. Suddenly, the primary skill required to "program" is not mastery of a specific coding language or a visual development environment, but rather the ability to articulate a problem and desired solution clearly. This opens up the possibility for nearly anyone with a business problem to conceive and even initiate the creation of a digital solution. The constraints on the types of applications that can be built are also diminishing, moving beyond simple data entry forms or basic automations to potentially more complex, interconnected systems.
Beyond the Hype: The Enduring Complexity of Business Logic
Despite the excitement surrounding generative AI’s potential to democratize software development, a critical analysis reveals that the core challenges of enterprise software—particularly those related to complex business rules and logic—are unlikely to be fully circumvented. Many proponents of the new user programming paradigm suggest that it could lead to a future where everyone builds their solutions, rendering traditional SaaS providers obsolete. However, this perspective often overlooks the deep-seated complexities embedded within mission-critical business applications.
Consider enterprise software categories such as procurement, invoicing, payments, budgeting, forecasting, payroll, staffing, sales force automation, customer relationship management (CRM), and customer service. These systems are not merely data repositories or simple workflow engines. They are intricate ecosystems designed to enforce literally thousands of often highly complex business rules and execute millions of lines of associated business logic. These rules dictate everything from regulatory compliance (e.g., GDPR, HIPAA, SOX), financial reporting standards (e.g., GAAP, IFRS), security protocols, legal frameworks, pricing structures, and operational policies.
The effort required to discern, codify, and manage these rules is immense. They are frequently a product of years, if not decades, of operational experience, legal precedents, and evolving market conditions. The vast majority of non-technical individuals, even those intimately familiar with their daily tasks, have little to no awareness of the intricate web of underlying business rules that govern their processes. Even experienced technical professionals often struggle to fully grasp these rules, which are frequently embedded deep within legacy code, sparsely documented, or understood only by individuals who have long since left the organization. This lack of comprehensive, accessible documentation is a significant hurdle, as evidenced by the challenges inherent in addressing technical debt—where untangling and re-evaluating existing business logic is a prerequisite for modernization.
Product managers and business analysts traditionally play a crucial role in this discovery process, meticulously working to uncover, define, and translate these nuanced business requirements into viable software solutions. A system that effectively captures, manages, and enforces these thousands of essential business rules, ensuring transactions are processed correctly and compliantly, represents significant intellectual property and development effort. This inherent complexity serves as a formidable barrier to entry for purely user-programmed solutions, especially in highly regulated or financially sensitive domains.
The Rise of Interoperability: Model Context Protocols
While the robust SaaS vendors managing these complex business rules are unlikely to disappear, the way their solutions are consumed and integrated is poised for a significant transformation. The future of enterprise software points towards a more interconnected and composable architecture, where purchased solutions are not just used by humans, but also by intelligent AI agents and custom-built workflows.
A major enabler for this shift is the emergence of widely accepted protocols designed to describe business services in a machine-readable and machine-understandable format. This need has existed since the dawn of the internet, but only recently has a viable solution gained significant traction. Anthropic’s proposal for the "Model Context Protocol" (MCP) is a notable example. MCP aims to provide a standardized way for AI models to understand the capabilities, inputs, and outputs of various software services, allowing them to interact with these services autonomously and intelligently.
This protocol represents a critical architectural advancement. By enabling AI agents and custom solutions to seamlessly access and control complex component services, it facilitates a new era of integration and automation. Instead of humans directly interacting with every SaaS interface, AI agents could act on their behalf, orchestrating workflows, extracting data, and executing tasks across multiple platforms.
The Hybrid Future: Synthesis of Built and Bought Solutions
The trajectory suggests that the future of the build-versus-buy decision will not be an either/or proposition, but rather a "yes to both" scenario, emphasizing intelligent integration and customization. Companies will continue to procure sophisticated and valuable component services from specialized SaaS vendors for core business functions. These bought components, however, will be designed with a new imperative: to be equally accessible and controllable by both human users and software agents.
This vision entails several key shifts:
-
AI Agent Orchestration: AI agents, potentially developed by the SaaS vendors themselves, by system integrators, or even by end customers, will act as intelligent intermediaries. They will interact with the purchased services, executing complex business processes, automating routine tasks, and providing insights that were previously difficult to obtain. For instance, an AI agent could manage a procurement process end-to-end, negotiating with vendors through an API, processing invoices via a financial SaaS, and updating inventory in an ERP system, all based on predefined business rules and real-time data.
-
Custom Workflows via Generative AI: Beyond AI agents, customers and system integrators will leverage generative AI tools to create custom, "vibe-coded" or hand-coded workflows that sit on top of these component services. These bespoke solutions will address highly specific business needs, integrate disparate systems, and provide unique competitive advantages, without requiring companies to build foundational infrastructure from scratch. This allows for agility and differentiation while still benefiting from the robustness and compliance of commercial offerings.
-
Enhanced Interoperability: The widespread adoption of protocols like MCP will standardize the way software components communicate, significantly reducing the friction and cost associated with integration. This will foster a more composable enterprise architecture, where businesses can mix and match best-of-breed solutions and connect them seamlessly, creating truly adaptive and resilient IT environments.
Navigating the New Landscape: Skills and Strategy
The increasing mainstream adoption of user programming, amplified by generative AI, is a net positive for countless individuals and organizations previously constrained by limited IT resources. It empowers business users to be more proactive in solving their own problems, fostering innovation and reducing reliance on overburdened IT departments.
However, as non-technical people venture beyond simple personal time-savers and into creating more complex business solutions, they will inevitably encounter the same challenges that product development teams have grappled with for decades. The most critical lesson is that the technical act of building and delivering a solution is often the easier part. The truly hard part, and the one that separates successful products from costly failures, is discovering the right solution to build. This involves:
- Deep Problem Understanding: Moving beyond surface-level symptoms to understand the root causes of business challenges.
- Stakeholder Alignment: Collaborating with various departments to ensure solutions meet diverse needs and comply with internal policies.
- Business Rule Elicitation: Meticulously uncovering, documenting, and validating the thousands of explicit and implicit rules that govern operations.
- Viability Assessment: Ensuring solutions are technically feasible, economically justifiable, and legally compliant.
- User Experience Design: Creating intuitive and effective interfaces that people (and AI agents) can easily use.
In this evolving landscape, the role of IT professionals, product managers, and business analysts will shift from purely execution-focused to more strategic and consultative. They will become crucial guides in helping citizen developers navigate the complexities of business rules, data governance, security, and scalability. The future of enterprise software is not a wholesale replacement of bought solutions with built ones, nor vice versa. It is a sophisticated synthesis, where intelligent automation, natural language programming, and robust interoperability protocols converge to create highly customized, adaptive, and efficient digital ecosystems, pushing the boundaries of what enterprises can achieve.