ISSaaSACTUALLYDEAD?

The numbers point towards it. Figma down 76%. HubSpot down 64%. Monday.com down 53%. Salesforce down 40%. Atlassian down 56%. The market is repricing an entire category.

The internet is packed with bold claims that SaaS is dead. And with reason behind them. AI agents are replacing entire tool stacks. One assistant that reads your email, manages your projects, automates your workflows, and connects to every platform you already use.

Consider what this means practically. No more paying for 8 subscriptions to do 8 different things. No more navigating through 8 different interfaces daily. No more managing 8 different logins, or reconciling data across 8 disconnected platforms.

Tools like OpenClaw have emerged as viral demonstrations of this shift. A single local AI assistant that operates across WhatsApp, Telegram, Slack, Discord, and email. It reads, responds, schedules, automates, and learns. Early adopters report saving hours weekly on routine tasks.

The IDC (International Data Corporation) predicts that by 2028, pure seat-based pricing will be obsolete, with 70% of software vendors refactoring their pricing strategies around new value metrics, such as consumption, outcomes, or organizational capability. The monthly “per user” payment method cannot be expected to survive when one AI agent is able to do the work of five seats.

If software is priced by the number of humans using it, and AI reduces the number of humans needed, the entire business model unravels.

So if all this is true, why are these predictions premature?

Humans are creatures of habit

Figma may be down 76%, but designers are not leaving. Years of muscle memory, custom templates, plugins and established team workflows create switching barriers that no AI agent can just overcome overnight. And the same applies to Salesforce, Workday, and others. More than just tools, they are embedded systems of record that organizations have spent years (and often large sums of money) implementing.

Depth still needs human judgement

AI can do a lot brilliantly. It can automate 80% of many workflows with impressive accuracy. But that remaining 10 to 20%? It requires understanding your specific client relationships, the people in your business and logic specific to your operations. And when the stakes are high, only systems built for your operation will suffice. Not a general purpose trying to interpret your intentions. Consider a complex HR workflow like employee onboarding. With traditional SaaS, this involves multiple steps and tools. Creating accounts, enabling access to tools, initiating payroll… An agentic AI system could theoretically automate much of this. But what happens when the new hire has a non-standard visa situation? Or when the role requires access to sensitive client data?

These edge cases, which are not really edge cases at all in a real organisation, require human judgment, institutional knowledge, and sometimes manual intervention. AI can take care of routine, yet humans are still needed.

This non withstanding, direction is clear. The debate is about how fast your software stack adapts to a world where AI agents execute tasks and humans focus on finding and solving problems.

Companies getting ahead now are not working for the perfect answer. They are already adapting their infrastructure, renegotiating their contracts, and building the custom systems that AI cannot easily replace.

The question is not whether SaaS will survive. The question is whether your business is positioned for what comes next.

Are you?