AI agents instead of SaaS. What will replace familiar corporate software

Investments in SaaS companies were once considered one of the most reliable avenues. Their business model was simple and clear: create a cloud platform for the widest possible audience, sell subscriptions based on the number of users, and scale the business through customer base growth.

This is how Salesforce, Workday, and many other companies grew, bringing billions of dollars to investors. However, the situation is changing rapidly today. Artificial intelligence is changing the very principle of software operation. AI agents are beginning to perform tasks instead of people, and the traditional per-user pricing model is gradually losing its relevance.

According to Richard de Silva, founder and managing partner of Lateral Investment Management, the SaaS market has already reached its peak. A new era is now beginning—the era of AI-native software, which automates the work of specialists and accesses a much larger market.

The end of the per-user pricing model

For many years, most SaaS products used a pricing model based on each company employee. But as soon as AI agents begin to perform a significant portion of tasks, this scheme stops working.

For example, if a company previously needed 100 CRM licenses for its sales department, in the future it might only need 50 licenses, because part of the functions will be taken over by artificial intelligence.

Therefore, software developers are looking for new ways to monetize. Instead of charging per user, they are starting to charge for completed work or achieved results.

For instance, a legal AI platform might charge for each contract prepared, effectively performing part of a lawyer's work. A corporate expense management service might take a percentage of the identified overspending. A platform for refunding erroneous payments might charge a commission only on successfully refunded funds.

The new wave of AI-native products is built not around employee collaboration, but around automating the actions of knowledge workers. These solutions are beginning to compete not only for companies' IT budgets but also for personnel budgets.

Universal SaaS services are becoming vulnerable

According to the author, universal SaaS products designed for the mass market face the greatest risks today.

If a service is merely a shell around a workflow that AI can perform independently, its value decreases rapidly.

The following are at risk:
— form builders
— project management platforms
— CRM systems for small businesses
— social media publishing automation services
— other universal tools.

The author believes that these product categories are already beginning to lose ground rapidly and may not recover.

Why niche AI companies win

In the new reality, specialized platforms operating in specific industries gain an advantage.

The author highlights three key advantages of such companies:

Long-term customer base. Ongoing relationships with clients allow for a better understanding of their tasks and deeper integration into daily business processes.

Deep industry expertise. Companies know the specifics of regulated industries, understand legislation, professional terminology, and the nuances of workflows.

Proprietary data. It is unique data that becomes the main competitive advantage. It is unavailable to universal AI models and allows for the creation of more accurate solutions.

When a product is built around the processes of a specific industry, switching providers becomes significantly more difficult. This is no longer just about data migration, but about restoring accumulated knowledge, settings, exceptions, and experience that have been built up over years.

For example, exporting a contact list from Salesforce is relatively easy. However, migrating the logic of bank scoring or an insurance risk assessment system is much more complex.

People remain part of the product

Despite the development of AI, the author does not believe that people will completely disappear from workflows.

The next generation of enterprise software will be built on the Human-in-the-Loop model—a combination of artificial intelligence and human expertise.

This is especially important for industries such as healthcare, law, construction, cybersecurity, financial services, and the defense industry.

Routine operations will be maximized automated, but the making of critically important decisions will remain with humans, as the cost of error is too high.

This approach changes the very nature of software. The development company becomes not just a software provider, but a full-fledged client partner involved in implementation, process configuration, quality control, and continuous product improvement.

Each new project helps the company accumulate industry expertise, improve models, and create an increasingly high barrier for competitors.

The market becomes significantly larger

According to McKinsey, artificial intelligence has the potential to generate up to $6 trillion in additional productivity annually in the global economy.

Therefore, AI-native platforms are beginning to compete not only for software budgets but also for personnel expenses, regulatory compliance, and risk management.

This is a significantly larger market than the one traditional SaaS companies operated in.

According to Richard de Silva, the winners will not be companies that simply add AI to existing products, nor those that create a service layer on top of SaaS.

Leaders will be companies with deep industry expertise, proprietary data, and AI-native architecture, which fully integrate software and services into a single system.

It is precisely such companies, the author believes, that will define the development of enterprise software in the coming decade.