Asana’s stock is down roughly 50% in the past year — and the story behind that slide is as much about AI as it is about Asana. Its problem? Being a subscription SaaS product in a world where a developer can prompt an AI to build a functional clone in an afternoon — for free.

Business Insider recently explored what some are calling the “SaaSpocalypse”: the accelerating trend of developers using AI coding tools to build custom replacements for standard business software, cutting out subscription fees entirely. It’s a real phenomenon, and it has direct implications for how businesses think about their software stack.

Key Takeaways

• AI tools like Claude Code and Cursor can now replicate simple SaaS tools in hours.
• A 20-person team spending $25/seat/month on project software could build a viable replacement.
• Simple tools (kanban boards, basic CRMs, dashboards) are most vulnerable to AI-built clones.
• Deep integrations, network effects, and regulated domains still protect enterprise SaaS.
• Every software procurement decision now has a credible “build” option worth evaluating.

What Can Vibe Coding Actually Produce?

The term “vibe coding” — popularized by AI researcher Andrej Karpathy in early 2025 — refers to describing what you want in plain language and letting an AI agent write the actual implementation. Tools like Claude Code, Cursor, or GitHub Copilot do the heavy lifting; you iterate by feel rather than by reading every line. As of April 2026, these tools can produce functional, deployable applications from natural-language prompts.

For a tool like Asana — at its core a task manager with assignments, due dates, and project views — the answer to “can AI replicate this?” is: a lot. A solo developer with a weekend and Claude Code can now produce something that handles the core workflows for a small team. It won’t have Asana’s integrations, enterprise SSO, reporting dashboards, or five years of UX polish. But it’ll handle the basics.

That’s enough to matter financially. A team of 20 paying $25/user/month is spending roughly $6,000/year on a tool that, functionally, can be replicated for the cost of a cloud server and some developer hours. For a broader look at how AI agents are reshaping how businesses operate, see our overview of AI agents in autonomous business workflows.

Where Does SaaS Still Win?

Asana’s actual counterargument is worth taking seriously. The company has pointed out that its real value isn’t task management — it’s coordinating large organizations, including AI agent workflows alongside human workers. At 500 people across 30 departments with cross-functional dependencies, version-controlled integrations, and executive reporting, a vibe-coded clone falls apart fast.

The SaaSpocalypse isn’t evenly distributed. The most vulnerable products are:

• Simple horizontal tools: Basic CRMs, kanban boards, form builders, simple analytics dashboards
• Tools where the core UI is the product: If the value is entirely in the interface and not in the data network or integrations, AI can replicate the surface layer quickly
• Low-switching-cost tools: Products where your data isn’t locked in and migration is easy

The more defensible products are:

• Deep integrations: Tools that are woven into dozens of other systems (Salesforce, Workday, etc.)
• Network effects: Products where value scales with how many people inside or outside your organization are using them
• Regulated domains: Healthcare records, financial compliance, legal documentation — domains where custom-built tools face significant validation burdens
• AI-native platforms: Tools that are themselves evolving to incorporate AI agents into their core workflow

How Does This Change the Build-or-Buy Decision?

For business leaders, the more important framing isn’t “will AI kill Asana” — it’s that every software procurement decision now has a viable “build” option that didn’t exist 18 months ago.

That changes the negotiating dynamic. When you can credibly say “we can build this in two weeks with Claude Code,” enterprise software vendors lose pricing power. We’re already seeing this play out in contract negotiations.

It also means internal technical teams are taking on more product surface area. That comes with maintenance obligations, security responsibilities, and feature backlog debt. A vibe-coded Asana clone has no security team reviewing pull requests. If you’re evaluating specific AI coding tools for your team, our 2026 AI coding tools roundup covers the leading options in detail.

What Should Business Leaders Do Now?

A practical framework for right now:

Audit your SaaS stack by complexity tier. Tools doing simple, well-defined jobs are the most replaceable. Tools handling complex workflows, deep integrations, or compliance are not.

Use AI to reduce SaaS sprawl, not eliminate all vendors. The goal isn’t to build everything in-house — it’s to stop paying for tools that no longer justify their cost given what AI can produce.

Don’t confuse a prototype with production software. Vibe-coded tools can bootstrap fast and handle internal workflows. They are not substitutes for enterprise-grade software in customer-facing or regulated contexts without rigorous review.

Watch for AI-native SaaS. The most durable software products right now are the ones incorporating AI agents as first-class features — not just adding a chatbot, but restructuring how work gets done inside the product. That’s a different competitive moat.

The Bottom Line

The SaaSpocalypse is real, but it’s selective. AI can now clone the surface of many SaaS tools quickly. That matters for pricing power, procurement decisions, and which software companies survive the next three years. But complexity, integrations, and network effects still win — for now. The question for every business is which side of that line your current tools sit on.

---

Frequently Asked Questions

What is vibe coding?

Vibe coding is a term coined by AI researcher Andrej Karpathy for the practice of describing what you want software to do in plain language and letting an AI tool write the actual code. Instead of knowing programming syntax, you explain the outcome you need, and tools like Claude Code, Cursor, or GitHub Copilot generate the implementation.

Can you really replace SaaS tools like Asana by building your own with AI?

For simple internal tools, yes. A 20-person team paying $25/user/month on Asana spends roughly $6,000 a year on something a developer can replicate in hours using AI coding tools. The catch: vibe-coded replacements lack enterprise integrations, compliance features, and the polish needed for customer-facing use. They work best as internal workflow tools.

Which SaaS products are most at risk from AI-assisted custom development?

Simple, single-purpose tools with no deep integrations or network effects are the most vulnerable — basic project trackers, form builders, internal dashboards, and lightweight CRMs. SaaS products that benefit from network effects, heavy third-party integrations, regulated industry requirements, or are AI-native platforms are much harder to replace with a custom build.

Should every company now consider building instead of buying software?

Every procurement decision now has a credible “build” option worth evaluating, which is new. But building still requires developer time, ongoing maintenance, and security oversight. The calculus favors building for simple internal tools with predictable requirements. It rarely makes sense for complex, customer-facing, or compliance-heavy software where enterprise-grade reliability matters.