"SaaS is dead because anyone can create an AI clone in a day"
I've seen a good few tweets recently to this tune and I wanted to give my thoughts on it.
Cloning a SaaS app has never been rocket science. For years now, Trello and Jira clones have been a common demo app for web frameworks to showcase their functionality. So then why would anyone pay for Trello or Jira? Because it's low-risk.
Yes, it's definitely easier to clone the basic feature set of a given SaaS app now with LLMs, but the basic feature set is only part of the story of why most companies pay for SaaS apps.
It's a boring answer, but every company decision to choose a specific vendor or build a solution in-house always boils down to the same risks and trade-offs. Things like
In one thread it was suggested that the SaaS market would be won by indie hackers creating low-cost AI-built alternatives to existing SaaS apps. Given the trade-offs above, it seems unlikely that a sizeable chunk of companies of any decent size are going to choose a one-man show SaaS for something business critical like task management, internal chat, or support instead of an off-the-shelf solution like Jira, Slack or Intercom respectively.
Others have suggested that companies would build and maintain these SaaS systems internally because it's now so easy, another idea which I don't think has fundamentally changed since the advent of LLMs. Building internally does not mean you get it for free. Once you've cloned the basic feature set, as the company grows you'll require things like SSO, permissions, new functionality, bugs to fix, and uptime to guarantee. This all comes with a tangible man-hour cost that will almost certainly be more than the cost of paying a SaaS vendor to handle these things for you.
There have always been mature open-source alternatives to many SaaS apps that companies already pay for. There are open-source self-hostable alternatives to Jira, Slack, GitHub, Stripe, Mailgun etc but companies prefer to pay an established vendor to build and manage it because it incurs so much less risk and time.
All this is to say the build vs buy SaaS equation has not changed. The real value isn’t in the initial build with the flashy UI, but in the reliability, maintenance and security that you get from established vendors. There were similar claims about the death of SaaS and the end of software development when “no-code” solutions came along and the wheels kept on turning because the hard parts are still just as hard.