The end of the SaaS bubble: GenAI is giving companies back control over software

The digital world is undergoing a revolution that many have not yet fully appreciated. Two seemingly independent forces are coming together to change the game for small and medium-sized software companies.

The first driver is the massive deployment of generative AI, which is dramatically changing the way applications are developed. Development is getting easier, faster and, most importantly, cheaper. The second force represents the dominance of SaaS models, which have become key tools for smaller and mid-sized companies over the past two decades, enabling them to get advanced functionality in exchange for regular monthly payments that are less than it would cost to develop software in-house.

What is SaaS?

SaaS (Software as a Service) is a model for delivering software applications where the software is hosted on the provider's servers and end users access it, for example, through a web browser. This model eliminates the need for local installation and management of the software, with all updates, maintenance and security provided by the service provider. SaaS allows organizations to flexibly scale application usage and minimize IT infrastructure costs.

But today, thanks to the explosion of generative AI, smaller and mid-sized businesses are gaining the ability to reach for more cost-effective customized solutions and get rid of (or consolidate) the overhead of paying for external SaaS. This is largely due to a few key factors.

Key factors positively influencing software pricing

1. Faster development due to great language models

Involving LLM in the software design and development process radically reduces the time it takes to build a working application. Tasks that used to take weeks can now be completed in days. Automatic generation of code, tests and documentation reduces development costs by up to 40-60%.

2. Smarter applications with lower logic costs

The advent of RAG (Retrieval-Augmented Generation) systems and MCP (Model Context Protocol) servers makes it possible to build advanced AI directly into application logic. Instead of complex business rule programming, we can use AI to solve complex problems with minimal coding.

3. Next-generation integration platforms

Tools like n8n and similar AI-integrated platforms make it possible to solve selected tasks without having to write specialized code. AI helps automate workflow and connect different systems in an intuitive way.

The evolution of the SaaS market

Current data confirms changes in the SaaS market, whose strength is declining while generative AI is becoming a means for faster and cheaper application development.

It is also significant that enterprises today manage on average a few dozen to hundreds of SaaS applications, with the IT department controlling only a fraction of the spend. According to a survey by US-based Zylo, organisations around the world are only using approximately half of their SaaS licenses, which represents a waste of financial resources that totals up to £500 million per year in unused licenses.

From success through the bubble to rationalization

The success of early SaaS (2000-2010)

The use of SaaS started around 2000 and gradually became the main form of software application deployment. Pioneers like Salesforce (2000), Google Apps (2006) or Czech success stories like Bohemian Coding (Sketch, later sold) showed the way. The SaaS model solved real problems for companies, such as no installations, automatic updates, availability from anywhere.

The bubble and the startup boom (2010-2020)

Success motivated thousands of new entrepreneurs. Cheap money and the pre-Covid startup boom created an environment where SaaS for every conceivable activity was created. Motivated IT enthusiasts founded projects at any cost. Some out of passion, others with visions of skyrocketing growth and getting rich quick. But the more astute ones understood that gold diggers are best sold shovels and rented saloons. And so an ecosystem of startup events, incubators, coworking spaces and business coaching began to grow in parallel.

Cooling and rationalisation (2020-present)

After Covid, the market began to rationalize. Rising prices, market corrections and the end of cheap money brought more cautious investors. But paradoxically, the number of new SaaS projects continues to rise - instead of robust investor projects, specialized tools from smaller companies are emerging.

There are a number of firms "spamming" LinkedIn and desperately trying to sell their SaaS to as many clients as possible. The market where the greatest oversaturation of these tools is occurring is marketing. Hundreds of external SaaS vendors or various interconnects are stuck on the largest e-commerce platforms. Yet in many cases, this is a software "hatchet soup".

The arrival of GenAI and the redefinition of the market (2023-present)

According to Stanford analysis, young workers in the most exposed AI occupations (including software engineering, marketing, and customer service) have seen a 6% decline in employment. In May 2023, 3,900 jobs were lost in the U.S. directly due to AI, making it the seventh largest job eliminator that month.

Many forward-thinking tech companies and investors are now hesitant to invest in new SaaS - the threat is that something will emerge during development that will outperform their product. At the same time, the need for companies to protect their know-how, data and code is rising, as AI is making it much easier to copy data or even entire software.

Where it's going: A return to custom solutions

In the words of Michal Rost and Mark Rost of TRITON IT, small and medium-sized businesses now have a unique opportunity to strengthen their digital vertical, get rid of vendor lock-ins and consolidate the confluence of external lump-sum payments for a plethora of SaaS.

Specific examples of transformation

Location: 5 different specialized B2B SaaS for 25,000 CZK per month
Solution: One custom application for 150,000 Kč one time and a significantly lower service fee

Benefits of custom solution:

• Full control over functionality
• No monthly fees
• Data protection and know-how
• Possibility of precise adaptation to company processes
• Independence from external suppliers

Technological possibilities of 2025

With tools such as:

• AI code generators (Claude Dev, GitHub Copilot, Cursor)
• RAG systems for intelligent corporate data processing
• No-code/Low-code platforms with AI assistance
• N8nintegration tools with AI capabilities

can create robust applications at less than half the cost of three years ago.

A specific example

Michal Rost from TRITON IT shows an example of how the digital vertical can be demonstrated in practice.

We dealt with a case involving a client who used a marketing agency to manage PPC campaigns. The client paid for the credits themselves and also a considerable agency fee. In addition, the agency recommended specialized SaaS for campaign automation, which the client had to pay for separately. However, after rebranding and a change in ownership structure, the monthly fee for this one tool increased to more than CZK 20,000. When we analyzed the situation, it turned out that the functionality of this tool could be fully replaced by our own scripts and a simple automation workflow. The client thus saved a considerable amount of money and gained control over both the data and the process.

Deliberate transformation instead of revolution

We are not witnessing the end of the SaaS era, but its transformation. Companies that can properly assess their needs and judiciously combine SaaS tools with custom solutions will gain a competitive advantage.

GenAI is indeed democratizing software development, but it's an evolution, not a revolution. The key to success remains technical expertise, i.e. the ability to decide what to develop, what to buy and how to bring it all together into a functional whole.