Artificial intelligence is reshaping every sector, particularly software. The integration of GenAI into multiple aspects of business operations or product offerings continues to drive new innovation and create new system design challenges.
Rishi Jaluria, software analyst at RBC Capital Markets, recently shared his thoughts on what’s top of mind for many investors thinking about the next chapter in software’s AI evolution.
Will AI agents replace software?
Over the long-term, we may see pre-packaged SaaS applications be replaced by agents. We would expect to see companies use specific agents for functions like sales, marketing, and HR, rather than the current architecture of bolting on agents to existing infrastructure. This does not mean the “death” of software, but rather a changing definition.
Before the internet, obtaining software required purchasing physical media, such as CD-ROMs or floppy disks, and manually installing it on a computer. Later, software was downloadable but still needed on-premises installation. Today, software typically refers to cloud-based services and SaaS applications accessible online through web and mobile browsers.
In the future, we expect the definition of “software” to refer largely to GenAI agents with autonomous decision-making and task execution, surpassing the functionalities of conventional software. We believe this would force existing SaaS companies to fully rearchitect for AI, although it may take several years, if not more than a decade.
AI drives strategic M&A
The rise of SaaS led to larger software vendors buying cloud upstarts to accelerate their roadmap to the cloud. Interestingly, many of these acquisitions happened after the 2008 Global Financial Crisis, as recessions can trigger waves of M&A.
Similarly, we believe larger software companies will likely acquire AI-native companies to accelerate their AI roadmaps or make major investments in them. We may see many of the AI-native unicorns or centaurs (companies with $100 million in annual recurring revenue), get acquired by large SaaS platforms in the coming year, especially if those AI-native companies could become potential competitors.
It is important to note that incumbents maintain some major advantages, including leveraging scale and distribution, balance sheets, and being able to buy proven companies.
“Crossing the AI chasm” is difficult
Building true GenAI and agentic platforms is difficult. While most software companies today are bolting AI features and capabilities onto traditional SaaS platforms, it requires major investment, time and innovation to truly rethink the platform for a post-AI world.
Truly innovating and integrating AI also presents challenges, including scalability, data silos and the need for cultural shifts within the organization (not to mention potential cannibalization risk).
In our view, there is a major difference between AI-native and AI-enabled, and the perceived incumbency and data advantage by large vendors may be overstated.
As a result, we believe it is critical for investors to scrutinize vendor AI strategies and product offerings and get beyond marketing to uncover the reality. Investors also need to keep a close eye on AI-native vendors that could turn into competitors or disintermediators over time.
Rishi Jaluria authored “Jaluria Software Download 21: Agents Replacing Software, GenAI M&A, and Crossing the AI Chasmatism,” published on March 12, 2025. For more information on the full report, please contact your RBC representative.
