After Adobe, Microsoft Office Is Next: AI Is Rewriting the Economics of Software

From GenOffice to WordCraft, GridCraft and DeckCraft — and what comes next

The next chapter after Adobe

In my previous article, I examined Brandon Thomas and his AI-assisted effort to recreate Adobe-style applications. The next chapter is already emerging: Microsoft Office alternatives. Two developments deserve attention. Genspark has released GenOffice, an AI-first, open-source office suite. Separately, Brandon Thomas has expanded his ArtCraft project into WordCraft, GridCraft and DeckCraft. They take different paths, but point to the same question: if AI sharply lowers the cost of building software, what happens to the business models of established vendors?

GenOffice: one engineer, one week — for an alpha

Genspark announced GenOffice in August 2026. Its team says one engineer built an initial alpha in one week, using about $10,000 in AI-model tokens. This is a striking claim, but it is not the total cost of a finished enterprise product. It excludes later engineering, security, support, infrastructure and long-term maintenance. GenOffice offers Docs, Sheets, Slides, PDF, Markdown and HTML editing. It opens and saves familiar .docx, .xlsx and .pptx files. Its distinguishing idea is an AI agent working inside the editor: revise a passage, construct formulas, or create editable slides. The project is available under an open-source license for Windows, macOS and Linux.

A different route: Brandon Thomas and ArtCraft

Thomas, whose Adobe-style applications I discussed earlier, has also introduced WordCraft, GridCraft and DeckCraft. These are intended as counterparts to Word, Excel and PowerPoint. He reports using Anthropic’s Claude to accelerate development. The projects are open source, but early in development. They should not yet be described as mature substitutes for Office. Nor should AI-assisted reimplementation be confused with proven copying of proprietary source code. Compatibility, design inspiration and reverse engineering are distinct issues, with legal boundaries that vary by method and jurisdiction.

How can ordinary users try GenOffice?

Start with the official GenOffice site or GitHub repository. Download the installer for your operating system, then open a non-sensitive copy of a Word document, spreadsheet or slide deck. Try a simple instruction: ‘Rewrite this introduction for senior executives,’ ‘Create a monthly summary using formulas,’ or ‘Turn these points into six editable slides.’ Review the proposed changes and compare the output with the original in Microsoft Office. GenOffice can connect to Genspark’s AI services or to an API key for a supported model. The application is free, but AI calls can consume credits or incur provider charges. Local document storage does not automatically mean every AI request stays private: check the selected model provider and data-handling settings before uploading confidential material.

What about other alternatives?

LibreOffice and ONLYOFFICE already demonstrate that office software need not be owned by Microsoft. Their existence is important context: the novelty is not simply a free Word alternative. It is the speed of AI-assisted development and the emergence of AI-native workflows. ONLYOFFICE supports document collaboration and AI integrations; LibreOffice remains an established, mature open-source suite with a different approach to AI. New entrants must compete on reliability, fidelity, collaboration and trust—not only on screenshots or rapid coding demonstrations.

Why Microsoft Office will not disappear overnight

Office is more than three desktop applications. Businesses depend on formatting fidelity, complex Excel models, macros, identity and access controls, document retention, accessibility, security, support and Microsoft 365 collaboration. A generated spreadsheet that looks convincing can still contain incorrect formulas. A document that opens may still break pagination or references. Enterprises therefore face migration and operational costs that are much larger than the price of a license. A promising alpha is evidence of changing development economics, not evidence that Microsoft’s enterprise moat has vanished.

What next? My view of 2027–2029

I see three stages. First, AI will enable many small teams to release credible prototypes in weeks rather than years. Second, some projects will mature into focused products that solve specific user problems better or more cheaply than broad suites. Third, established vendors will face pressure to simplify products, improve AI functionality and justify recurring prices. These are scenarios, not guaranteed timelines. The winners may not be the teams that clone the most features. They may be the teams that rebuild a workflow around what users actually need.

The larger opportunity: replace legacy workflows, not just interfaces

This is especially relevant to banking and large enterprises. Many organizations still operate expensive, fragmented systems designed for older infrastructure and processes. AI-assisted engineering may allow teams to rebuild selected workflows with modern APIs, cloud architecture and embedded agents. Consider branch servicing, onboarding, reconciliations, internal reporting or case management. The opportunity is significant—but the standards are higher. Transaction accuracy, audit trails, cyber controls, regulatory obligations and resilience cannot be delegated to a coding model. Human engineering judgment and disciplined validation remain essential.

The economic question

If the cost of creating a first working version falls dramatically, software supply could expand. More developers, startups and internal technology teams may enter markets once dominated by large vendors. That could lower barriers, stimulate competition and shift value away from routine coding toward distribution, domain expertise, proprietary data, integration and trusted operations. It could also create a flood of poorly maintained applications. Lower development cost does not eliminate the cost of owning software.

And then what next?

My conclusion is not that Microsoft Office is finished. It is that software creation is becoming more accessible. GenOffice and ArtCraft are early signals of a structural change: users may increasingly ask why an application is costly, complex or inflexible when a small team can build a simpler alternative. Over the next three years, the critical question for every software vendor will be: what value do we deliver beyond the code itself? For enterprises, the question is equally important: which legacy functions should we buy, which should we modernize, and which can we now build? The next software revolution may be less about copying yesterday’s products and more about redesigning tomorrow’s work.

Sources and further reading

Genspark, GenOffice announcement (3 August 2026) https://www.genspark.ai/blog/genoffice-open-source-ai-office

GenOffice official download and pricing notes https://genoffice.ai/

GenOffice source code and documentation https://github.com/genspark-ai/genoffice

Tom’s Hardware, Brandon Thomas’s Office projects (October 2026) https://www.tomshardware.com/software/microsoft-office/creator-of-free-adobe-clones-unveils-open-source-microsoft-office-excel-powerpoint-and-word-replications-autodesk-tools-and-other-subscription-based-services-may-come-next

ONLYOFFICE, self-hosted office suites https://www.onlyoffice.com/blog/2026/07/best-self-hosted-office-suites

Publication note: The forecasts and assessments are the author’s analysis. Product capabilities and maturity reflect public information available in October 2026.