The Coming Software Reckoning: When AI Can Rebuild What Customers Have Been Paying For
The future of SaaS, the rise of vibe coding, Adobe’s market dominance, and the strategic threat to established software companies
Introduction: The threat is no longer theoretical
For decades, the software industry has operated on a relatively stable assumption: developing sophisticated applications requires substantial investment, specialized engineering teams, years of accumulated intellectual property, and continuous maintenance. These barriers have allowed established software companies to build powerful market positions, sustain premium pricing, and maintain considerable control over their customers.
Artificial intelligence is changing not only how software is used but, more fundamentally, who can build it, how quickly it can be built, and what customers should reasonably be expected to pay for it.
The rise of vibe coding, AI-assisted professional software engineering, and increasingly autonomous coding agents introduces a competitive threat that many established software companies may still be underestimating. The threat is not simply that AI will make their developers more productive. It is that AI could make some of their products dramatically easier for competitors, customers, and small independent teams to reproduce, replace, or bypass.
In December 2024, I published The Future of SaaS in the Era of AI Agents, examining how intelligent agents might challenge the traditional application model. In November 2025, I completed another research study examining Adobe's long-standing shortcomings in supporting Arabic across its creative, document, and collaboration products. That research was not published at the time, for reasons I prefer not to disclose.
What happens when customers who have tolerated expensive, restrictive, or inadequate software for years finally acquire the technological means to build credible alternatives themselves?
1. The Adobe cloning story: A warning from an unfinished experiment
In October 2026, a development project targeting Adobe's Creative Cloud ecosystem attracted considerable attention. On October 7, Ars Technica reported that an AI-assisted developer was building open-source alternatives intended to reproduce the experience of familiar Adobe applications. The project drew attention because increasingly capable AI coding systems could dramatically reduce the effort required to recreate established software products.
The emerging alternatives have not demonstrated full feature parity with Adobe's mature professional products. Reproducing a recognizable interface is one thing. Reproducing decades of engineering refinements, complex file compatibility, advanced rendering, color management, reliability, and production workflows is another.
Nevertheless, dismissing the project because it is unfinished would be a serious strategic mistake. Its significance is not that Adobe has already been replaced. Its significance is that the economics of attempting such a replacement have changed.
The central question is not whether one developer can recreate the entire Adobe suite today. It is whether thousands of developers using increasingly capable AI systems can collectively erode the competitive advantages established vendors accumulated over decades.
2. From the agentic SaaS thesis of 2024 to the coding revolution of 2026
In my December 2024 article, I examined Microsoft CEO Satya Nadella's proposition that AI agents could increasingly replace conventional application interfaces as the primary way users interact with enterprise systems. I questioned whether that would eliminate SaaS altogether, concluding that software services would not necessarily disappear, but their architecture, delivery models, and sources of differentiation could change substantially.
The developments of 2026 introduce a second dimension: AI is no longer challenging only how customers consume software. It is challenging how software itself is produced.
Agentic consumption means AI agents execute tasks across systems, reducing dependence on traditional interfaces. AI-assisted production means developers and customers use AI to build alternatives, extensions, and customized applications. Together they pressure subscription pricing, differentiation, retention, and vendor-controlled workflows.
An AI agent might eliminate the need to navigate ten screens to complete a task. An AI coding agent might eliminate the need to purchase a separate application to perform that task in the first place.
3. Vibe coding is only the beginning
Vibe coding describes an approach in which a person describes desired software behavior in natural language and relies heavily on AI to generate and modify code. It lowers the barrier to producing prototypes and lightweight applications.
But the larger threat does not come from inexperienced users producing experiments. It comes from combining AI coding capabilities with professional engineering expertise. Experienced developers understand architecture, performance, security, databases, testing, deployment, and maintainability. AI can amplify those capabilities.
Vibe coding can expand who creates software; disciplined AI-assisted professional engineering can reduce the effort required to build and maintain credible alternatives.
Independent research, including METR's work on coding agents, illustrates both promise and limitations: agents can perform increasingly substantial software tasks, but reliability, judgment, and performance vary. The technical barriers protecting established software products are becoming less predictable.
4. Adobe’s subscription model: When customer dependence becomes a vulnerability
Adobe's competitive strengths are real: deeply embedded creative workflows, familiar formats, training ecosystems, third-party integrations, and accumulated user expertise.
But restrictive commercial practices can create strategic vulnerabilities. In June 2024, U.S. authorities pursued legal action alleging inadequate disclosure of certain subscription termination fees and difficult cancellation procedures. In March 2026, Adobe agreed to a settlement involving $75 million in civil penalties and $75 million in free services for eligible customers; Adobe denied wrongdoing.
This case does not establish that Adobe holds an unlawful monopoly. Market dominance, customer lock-in, and illegal monopolization are distinct concepts.
When customers remain because they value a product, that is competitive strength. When they remain because switching is difficult, the vendor may be accumulating dissatisfaction alongside recurring revenue. Customer lock-in can protect revenue temporarily. It cannot guarantee long-term relevance.
5. The Arabic-language experience: A case study in neglected needs
My November 2025 research examined Arabic support across Adobe's creative, document, and collaboration products. Throughout my career, I have contributed to regional software localization initiatives involving the Horizon Library System, Elluminate virtual classrooms, Moodle LMS, BigBlueButton, and other enterprise and educational applications.
Localization is not merely translating menus or adding fonts. It requires understanding how people work, communicate, read, write, teach, learn, and interact with technology in their linguistic and cultural environments.
The research documented historical problems including disconnected or reversed Arabic characters in certain workflows; inconsistent right-to-left formatting; long-standing complaints about Arabic OCR in Acrobat; difficulties editing and extracting Arabic PDF text; incomplete interface localization; and uneven Arabic interaction across collaboration and AI features. These issues varied by application and changed over time. Adobe also made meaningful improvements, including unified text engines in Photoshop and After Effects and sophisticated Arabic typography in InDesign.
On November 19, 2025, Adobe and Saudi Arabia's HUMAIN announced a strategic partnership focused on culturally aware generative AI, Arabic-first intelligence, ALLAM, Firefly Foundry, and potential integration across Adobe applications. On August 31, 2026, Adobe, HUMAIN, and Saudi Arabia's Ministry of Communications and Information Technology announced an expanded collaboration.
These initiatives are promising, but Arabic-first generative AI does not automatically resolve historical issues involving RTL interfaces, PDF editing, OCR, accessibility, or collaboration. Every unresolved customer complaint is now a potential product specification for an AI-enabled competitor.
6. The end of feature-based competitive protection
Traditional software companies have often differentiated themselves through feature breadth. Reproducing hundreds of features once required years of engineering. AI-assisted development puts pressure on this model.
Basic workflow screens, dashboards, forms, approval processes, reporting interfaces, and common integrations may be easier to reproduce. Advanced media processing, secure financial transactions, regulated decision-making, and large-scale enterprise operations remain harder.
Software companies must distinguish between functionality that was expensive to develop and value that is genuinely difficult to replicate. Customers do not necessarily need an alternative with 100% feature parity. They need one that delivers sufficient functionality, reliability, and economic value for their actual work.
7. The hidden threat: Customers becoming their own software providers
The conventional enterprise decision was framed as build versus buy. Purchasing established software was often preferable because internal development required high upfront costs, lengthy implementation cycles, specialized resources, and ongoing maintenance.
AI changes the assumptions underlying that decision. Organizations may increasingly develop internal dashboards, reporting tools, document-processing utilities, lightweight workflows, departmental knowledge portals, and specialized agents.
This does not mean every company should build its own ERP, CRM, LMS, or financial platform. Mature systems may remain safer and more economical. But the threshold at which internal development becomes attractive is changing. The emerging competitor may already be sitting inside the customer's organization.
8. What becomes of traditional SaaS economics?
The implications extend to pricing, margins, acquisition, and recurring revenue. AI-native competition can reduce development costs for repeatable tasks, move differentiation toward data and domain expertise, increase interest in consumption or outcome-oriented pricing, and shift user interaction toward agents.
A software company's price cannot be justified solely by its original development cost. It must reflect the economic benefit delivered compared with available alternatives.
The objective is not to abandon subscriptions indiscriminately. Many customers value predictable fees, and vendors need sustainable revenue for security, support, and innovation. The objective is to ensure the subscription represents value customers genuinely want to renew.
9. The dark future facing companies that fail to adapt
The danger is not necessarily sudden extinction. It is gradual erosion of relevance that conventional financial reporting may initially miss. A company could report strong recurring revenue while losing differentiation; retain customers on multiyear contracts while they experiment with AI-built alternatives; and continue releasing features while rivals focus on actual outcomes.
Possible futures include gradual commoditization of common functions; agent-driven disintermediation of application interfaces; selective replacement of peripheral tools; AI-native challengers outperforming legacy architectures; and successful reinvention by incumbents that use their resources, relationships, and operational expertise to adapt.
These are scenarios, not forecasts. The future is not predetermined. AI can strengthen incumbents as well as threaten them.
10. Why the future is not simply the death of software
Enterprise software requires far more than generated source code: security, performance, scalability, integration, accessibility, governance, privacy, compliance, disaster recovery, support, and maintenance.
Google's 2025 DORA research illustrates the complexity: AI adoption was widespread and many respondents reported productivity benefits, but organizational capabilities and software delivery quality still mattered. Faster development did not automatically guarantee stability.
AI reduces the cost of building some functionality but not necessarily the cost of operating it responsibly. Established vendors can build enduring advantages around reliability, trusted operations, proprietary data, regulatory expertise, and continuous customer value.
11. What software companies must do now
Conduct a ruthless product-value assessment. Identify what customers truly value, what is reproducible, and where dissatisfaction is accumulating.
Become AI-native rather than merely AI-enabled. Invest in agent-accessible APIs, modular architectures, structured data, intelligent workflows, and human oversight.
Replace artificial lock-in with earned loyalty. Compete through product quality, transparent terms, interoperability, and measurable outcomes.
Treat localization and accessibility as strategic differentiators. Build for regional languages, cultural requirements, and real workflows.
Reconsider the product portfolio. Modernize, consolidate, expose as services, or retire products where appropriate.
Transform engineering and product organizations. Equip teams to supervise coding agents, validate outputs, manage security, and rethink value propositions.
Redesign partnerships and ecosystems. Prioritize open integration, specialized expertise, and differentiated customer outcomes over simply expanding product catalogues.
12. A strategic test for software executives
Could a small team using current AI coding tools reproduce the most valuable 20% of our product?
How much revenue depends on switching costs rather than customer satisfaction?
Can customers and third-party agents securely access our capabilities through APIs?
Which unresolved complaints could become the basis of a focused competitor?
Are we investing in proprietary data, expertise, reliability, and outcomes that remain difficult to reproduce?
Does pricing reflect measurable customer value or historical market power?
Do product teams use AI to rethink the offering, rather than merely accelerate the existing roadmap?
Are we prepared for customers to build some functionality they currently purchase?
13. The opportunity for the Arab technology industry
Historically, many organizations in the Arab region have relied on imported platforms requiring localization, customization, and integration. AI-assisted development could change that relationship.
Local companies may develop specialized applications for Arabic-language workflows, educational systems, cultural heritage, institutional knowledge management, government services, and sector-specific enterprise operations.
Regional technology companies should not automatically assume their role is limited to reselling or customizing imported software. The opportunity to create globally competitive intellectual property may be expanding. But seizing it requires serious investment in engineering quality, domain expertise, security, sustainable commercial models, and long-term product ownership.
14. A new definition of competitive advantage
For much of the industry's history, competitive advantage came from possessing something others could not easily build: a sophisticated codebase, proprietary application, complex interface, or expensive engineering capability. In the emerging AI era, these advantages may become less durable.
The more important question may be what a company possesses that AI cannot easily reconstruct from publicly available knowledge and common software patterns: trusted relationships, specialized domain knowledge, unique datasets, validated workflows, operational reliability, regulatory expertise, or an ecosystem delivering superior outcomes.
Software functionality alone may become an increasingly weak foundation for long-term market power.
Conclusion: The real danger is strategic complacency
My 2024 SaaS article asked whether intelligent agents would change how software is consumed. My 2025 Adobe research asked whether a globally influential software vendor had responded adequately to the needs of Arabic-speaking users. The October 2026 cloning story connects these questions.
The first AI-generated alternatives may be incomplete. Some may fail. Others may never reach professional quality. But the cost of experimentation, development, and competition is changing.
The software industry is not necessarily approaching its end. It is approaching a period in which the economic value of software will be reassessed.
The greatest threat to established software companies may not be artificial intelligence itself. It may be their failure to understand what AI is doing to the economics of their industry.
The future will not necessarily belong to those who built the most software. It will belong to those who continue delivering the most value when software becomes easier to build.
Research references and further reading
- Usamah AlBaik, The Future of SaaS in the Era of AI Agents (December 2024).
- Usamah AlBaik, Adobe’s Arabic Language Support (unpublished research, November 2025).
- Ars Technica: AI developer takes aim at Adobe with open-source clones (October 2026).
- Google DORA: State of AI-Assisted Software Development (2025).
- METR Frontier Risk Report.
- U.S. Department of Justice: Adobe subscription settlement (2026).
- Adobe–HUMAIN Strategic Partnership (2025).
- Adobe, MCIT and HUMAIN expand partnership (2026).
Editorial note: This article combines documented developments, previously authored research, and forward-looking strategic analysis. Its competitive scenarios are possibilities, not established forecasts. The reported Adobe alternatives remain unfinished and are not proven replacements for the complete Creative Cloud suite.
#FutureOfSaaS #AgenticAI #VibeCoding #SoftwareIndustry #DigitalTransformation #ProductStrategy #Adobe #ArtificialIntelligence
Originally published in the VerbionPulse newsletter on LinkedIn: The Coming Software Reckoning: When AI Can Rebuild What Customers Have Been Paying For.
Comments