What web businesses will continue to make money post AI? If you can code practically anything with Claude Code (or equivalents), what types of web businesses will continue to stay viable or profitable? Source: AI is Killing Saas - https://nmn.gl/blog/ai-killing-b2b-saas |
We are a team of 4 people company and we use 15+ SaaS Web tools and we can code few of them but we see no reason to solve. Why would we replace Calendly already at just $12 per month? Or why would we will create an internal Outlook?
I assume expensive SaaS is generally very complicated nothing you can vibecode in a weekend, but I assume if someone is selling $50 SaaS, someone else will vibe code it and sell for $20
Or someone will sell calendly kind of app $10 for a year.
Or maybe everyone will rely on popular brands and this won’t make any meaningful difference in the market other than few categories.
SaaS was always about the service part and works better when other companies use the same tool.
Unique software specific to business needs will get more interesting though.
This is usually masked by vertical integrate and buyout of competition.
So either your servicd raises prces to avoid buyout or it gets bought.
My current answer: go vertical and messy.
Ex, Healthcare portals with ugly data. Compliance platforms with painful regulations. B2B tools with 6-month sales cycles.
Source: I sell a complex B2B SAAS and even though customers are open to learning more about using AI and are curious, I never had a customer ask so far "Hey so now this AI thing is out, do I really need to you the $xx,xxxx/Year because I can build this in a weekend now, right?". The software is just 1 thing. Customers want stability, support, maintenance, someone to call if shit breaks and if a SAAS vendor does a decent job at it, they won't leave.
If anything, customers who are AI adaptors are asking to collaborate with their existing vendors if possible.
The value is in being able to sell it, being able to offset responsibility, all that ancillary stuff, which AI can't do at the moment.
Meta (Instagram, Threads), Bloomberg, X, YouTube, Snap, Netflix, TikTok, Valve.
Coding agents are not designed to clone network effects nor can they.
Most companies want to focus on their actual business, which is why they go to SaaS in the first place. They won’t want to do all this stuff themselves, as it’s a distraction from the core business, and they won’t do it as well as people dedicated to building a SaaS product.
This take is so unimaginably stupid and far from the truth it makes me angry. You have network effects, liability, maintenance, mental/managerial complexity, integrations and on and on the list goes. I would be weary of anyone proposing an issue and selling you the solution right with it.
You can also cook your own food. But do you? Do you break your own bread? Do you make your own pasta? No you don't. You pay for an outcome. SaaS is the same. Just because you can do it yourself doesn't mean you should. The code is the smallest part of a SaaS. This will be the same thing like "let's move devs to a cheaper country". Look how that turned out.
I think the point is that AI will be able to do this better and cheaper than the SaaS companies.
FIRESHIP: 7 AI updates breaking SaaS right now... https://youtu.be/cxcb55zr2Q8?si=D0OUeejL0d0SkXJZ
Some hypotheses
(A) AI helps most setup simple tools - even non technical people. But once you get beyond greenfield the cognitive debt builds so you can’t reason about it it. It’s possible mature products don’t get the same gains with AI (or have different types of productivity gains).
(B) We pay a SaaS company to be responsible for an SLA. I personally don’t want to be responsible for that SLA on my vibe coded app so I outsource it. See also support, etc
(C) We pay SaaS to be a reliable source of truth (like Shopify for my Ecom business). The app holds the state of something important. That investment in the ecosystem is itself a moat.
(D) Many “SaaS” businesses are not pure software. They handle payment, benefits, payroll. Often with complex human b2b backends. It looks like just software to us, but we pay to turn a complex set of human relationships into a slick dashboard.
If my business is a pure software tool, it doesn’t have a good moat - and frankly probably nevet did.
Real businesses aren’t going to over optimize on something that cost $10 per seat.
If it’s “always proven wrong “, is this going to be “The year of Linux on the Desktop”? Is “the iPad going to fail because it doesn’t run Flash?”? Are we going to start browsing the internet on our TVs using set top boxes to get on the “Information Super highway”?
Oh yeah and let’s not forget that the iPod had “less space than the Nomad and no wireless” and was “lame”.
Or don’t forget that Dropbox was going to fail because “For a Linux user, you can already build such a system yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem. From Windows or Mac, this FTP account could be accessed through built-in software.”
I can go on about Dvorak’s comment in 1984 “The Macintosh uses an experimental pointing device called a 'mouse'. There is no evidence that people want to use these things”
Or Michael Dell’s infamous “Apple should just shut down the company and give th money back to shareholders”
The IPod won because it had the Apple app store. People don't need the app store to acquire high quality content any more and they certainly don't want their content locked into the limitations of leased access to some distribution model.
> Self-hosted applications with a cloud premium upgrade are the new business model
That’s not happening for any company. Where pray tell are these companies “self hosting” and supporting their business that they are getting a better deal than their per seat licensing cost? Who is supporting their systems?
> For example why bother with all the invasive nonsense of Facebook if a person can self-host something that works better and share it with their real life contacts at their discretion?
So everyone is going to be a server admin and host software and set up their NAT correctly so they can access it via their home internet when their on their phones using their 30 Mpbd upload speeds from their cable modem?
Are people going to start running their own servers now?
> The IPod won because it had the Apple app store
That’s also not true. The iPod was released in 2001, the iTunes Music Store wasn’t released until 2003. As late as 2007 Steve Jobs said that only 5% of music on iPods came from iTunes in his famous “Thoughts on Music”