Is SaaS dying?

The evidence is more complicated than the headline.

Salesforce reported $41.5 billion in revenue for fiscal 2026, up 10% year over year. One company’s growth does not represent the entire SaaS market, and revenue growth does not prove every product or pricing model is durable. It does, however, make a blanket claim that established cloud software has already become obsolete hard to defend. Salesforce: fiscal 2026 results, February 25, 2026 ↗

Our view is that AI changes the basis of competition. Some narrow features may become easier to reproduce. Some products may be absorbed into larger platforms or replaced by custom workflows. Others may become more useful because an AI interface makes their underlying capabilities easier to access. These are different outcomes, not one industry-wide verdict.

A simpler interface does not remove the underlying work.

Consider an assistant that updates a sales opportunity after a meeting. Someone still needs to define which customer record is correct, which fields may change, who has permission, and what happens when the update is wrong. The assistant may reduce manual interaction while continuing to rely on the CRM underneath it.

This distinction matters when evaluating a replacement. A convincing demo can show the successful path. A business needs to understand the unsuccessful paths as well: missing data, conflicting records, unavailable services, and actions that should never have been allowed. Ask who owns those situations.

Compare the cost of the completed task.

A seat price alone is an incomplete comparison. An AI-enabled workflow may introduce usage charges, review effort, or additional integration work. A custom application may reduce subscription expense while creating maintenance and support responsibilities. A familiar SaaS product may still be expensive if people spend hours working around it.

Build the comparison around a real task. Include the software bill, setup, ongoing ownership, error handling, and the time people spend using it. Test with representative work rather than the easiest example. The goal is to understand the full cost of getting an acceptable result.

Choose what to buy, build, or connect.

Buying can make sense when an existing product handles a common need well and its operating model suits your team. Building can make sense when the workflow is distinctive enough to justify ongoing ownership. Connecting systems can be the most practical option when the tools are useful but the handoffs are not.

For example, a marketing team might keep its CRM, improve the corporate website, and build an integration that passes qualified enquiries into the right process. That can address a specific gap without rebuilding the entire customer system. It is an option to evaluate against the actual requirements, not a default prescription.

Ask vendors—and your own team—harder questions.

Before committing, ask how data can be exported, how access is controlled, and which functions are available through an API. Understand how changes are recorded and how a failed action can be corrected. For AI features, clarify what data is used and where human review fits.

Then ask who will own the workflow internally. Even a well-designed platform needs someone to maintain definitions, resolve exceptions, and decide what should change. A product that fits the organization is often more valuable than a longer list of features.

SaaS is being reconsidered at the level of tasks, interfaces, and value. Evaluate each tool by the work it supports and the responsibilities it takes on—not by whether it belongs to a category declared dead.