More than 100 public software companies are getting 'squeezed' by AI, according to a new study

• Generative AI is reshaping enterprise software, challenging traditional SaaS models.
• AI agents are evolving from assistants to core applications, impacting mid-market companies.
• Mid-size SaaS vendors face pressure from AI-native startups and tech giants like Microsoft, AlixPartners says.

A foundational shift is underway in enterprise software, and it's being driven by generative AI.

The rise of reasoning models and AI agents is beginning to erode the core assumptions that have defined the software-as-a-service business model for decades, according to a new study released on Monday by AlixPartners.

The consulting firm warned that this is squeezing more than 100 mid-market software companies, which are stuck in the midst of a powerful trend.

These companies are caught in a "big squeeze," pressured on one side by nimble, AI-native entrants that can replicate applications at a fraction of the cost and on the other side by tech behemoths, such as Microsoft and Salesforce, that are pouring billions of dollars into the AI arms race, AlixPartners said.

"We believe many mid-size enterprise software companies will face threats to their survival over the next 24 months," the firm added ominously. It declined to identify specific companies, given the sensitive nature of its findings.

AI: the new foundation, not just a feature

The most mature uses of AI in enterprise software today include copilots for software coding, such as GitHub Copilot from Microsoft, and support chatbots like Zendesk's Answer Bot. But these could be just the beginning. Generative AI is advancing from narrow use cases to the broader "logic and presentation layers" of software, the very foundation that traditional SaaS tools are built on, AlixPartners explained in the study.

This means AI agents are no longer just assistants within applications; they are becoming the applications themselves. These agents are capable of handling complex tasks, such as scheduling meetings, analyzing reports, and creating code, with little need for a graphical interface or structured workflow. And because they can run across various data types without needing extensive data normalization, they could render some traditional SaaS layers redundant.

"This shift could eliminate the need for many enterprise software companies that thrived in the traditional SaaS architecture," AlixPartners said.

A new threat from both sides

This puts mid-size SaaS vendors in a tricky position, the firm explained, citing a recent analysis it conducted of 122 publicly listed enterprise software companies with annual revenue below $10 billion.

It found that sales growth has slowed considerably lately. For instance, the percentage of high-growth companies in this group decreased from 57% in 2023 to 39% in 2024. This year, industry analysts are expecting further declines, indicating that only 27% of companies will be in the high-growth category.

AlixPartners also highlighted that software customers are moving around more than before. The median net dollar retention rate of enterprise software companies dropped from 120% in 2021 to 108% in the third quarter of 2024, the firm noted, citing data from Bank of America. (NDR is a common way to measure customer stickiness. When it's above 100%, that indicates revenue from existing customers is growing, while an NDR below 100% suggests revenue is declining from these sources.)

Many of these companies are now being undercut by AI-powered challengers with lower costs and faster iteration cycles. Simultaneously, larger players are integrating AI into their broader platforms, offering bundled functionality at lower price points through economies of scale, according to the consulting firm.

Klarna's recent decision to drop Salesforce and Workday in favor of smaller AI-powered vendors and in-house agents is a sign of where this trend may be headed, AlixPartners noted.

The SaaS model: ripe for reinvention

Traditional SaaS depends heavily on the user interface, structured data workflows, and seat-based pricing. But AI agents don't need dashboards, and they can function without rigid data hierarchies. This calls into question the relevance of the SaaS model itself, according to AlixPartners.

Some companies are already pivoting. Salesforce and ServiceNow have begun experimenting with outcome-based pricing for AI agents, where fees are tied to results, not user counts.

Among those 122 mid-sized software companies, AlixPartners found that half expect significant changes to business models in the next year.

At the same time, the compute costs associated with running AI agents can be significantly higher than for classic SaaS tools, making profit margin compression a potential threat. Software providers may need to rethink infrastructure strategies, possibly shifting to more efficient inference architectures, industry experts have warned recently.

Meanwhile, higher interest rates and tightening capital markets in recent years have put the onus on SaaS profitability, not just growth. Software companies have responded by cutting costs, optimizing portfolios, and rethinking pricing strategies.

According to the AlixPartners report, more than 60% of executives are now focused on AI as a growth driver. Unlocking that growth requires more than just product tweaks, it may require transformation across operations, go-to-market models, and customer relationships, the consulting firm suggested.

How to survive: a new playbook

So what's the path forward? The report discusses several strategic imperatives. Here are a few:

• Build AI agents: Not as bolt-on features, but as core products.
• Transform the business model: Move beyond a seat-based fee structure to usage- or outcome-based pricing.
• Streamline and focus: Shed low-growth products and reallocate R&D to AI development.
• Lean into M&A: For some, the best route may be to get acquired or consolidate.

In a world where generative AI tools can write or replicate software pretty well, differentiation must come from speed, relevance, and efficiency, not UI design or legacy feature sets.

The software era isn't ending. But the SaaS era, as we know it, is evolving. The next generation of enterprise tools may not be applications, they could be agents. And only the most adaptable companies will make the leap, according to AlixPartners.