Pull out your sackcloth and ashes. The SaaSpocalypse is upon us.
The doomsayer was the launch of plugins for Anthropic’s Cowork. If you’re unfamiliar with Cowork, it’s essentially a fancy wrapper for Claude Code, so that non-techies can take full advantage of the vibecoding revolution without diving into developer environments, too. Well, non-techies who use Mac. I’m on Windows, so I’m shut out for the time being. But I’ve watched a few demos, and my company has a similar internal tool that I recently used to fix a mandatory security update in a code repo I didn’t realize I owned, can’t interpret to save my life, and would never in a million years been able to fix manually.
Handy, that.
Cowork — not to be confused with Microsoft’s Copilot — now comes preloaded with a set of professional-assistant plugins that can perform specialized tasks such as contract review and competitive analysis at the mere click of a button, working directly with the files and tools you give it access to, such as your CRM system. All you need to do is share context about your business and its processes and preferences and, voilé, Cowork will complete the task in a fraction of the time, ready for you to go forth and conquer.
To further demonstrate how the plugins can democratize formerly specialized tasks, they’ve got labels like Legal and Marketing. As in the Legal and Marketing departments we no longer need because Cowork handles it.
Thus, the collective gasp you’re hearing is the rest of the professional world joining the existential crisis that folks who deal in content and coding have been suffering for the past year or so. It’s also the SaaS (software as a service) companies realizing no one needs to buy a soup-to-nuts platform anymore because they can use Claude Code to build exactly what they need. Without paying license fees.
Investors saw the same writing on the wall and started selling. It wasn’t a bubble burst, but was still enough to give the markets a pretty good rattle.
For those of you joining us in the existential crisis, welcome. We have hyperventilation bags. And cookies.
Hold your horse(men)
Seriously, though, it’s premature to panic. Yes, the world of knowledge work and professional services is changing and will rely heavily on agents to perform tasks we formerly asked humans to do. But in the frenzy over the SaaSpocalypse, ye olde humans are overlooking three key dynamics.
First, what Cowork’s plugins are doing was already fully doable. You could already build and train specialty agents, analyze with deep-thinking models, set up complex workflows of agents working together, build prototypes, and automate all manner of business tasks in your other tools. I’ve done this myself, in fact, as part of a class I took last year. The caveat is you often had to be a bit techie and tool-savvy to get the result you were looking for. Knowing about APIs and MCP servers, for example, or which tools work better for which tasks.
The Cowork plugins took off like a rocket because they skip over the techie stuff and go straight to “let me do that for you.” But I guarantee a lot of companies already had their tech teams working on tools like this. And I guarantee the SaaS companies already knew it. There’s nothing new here, except perhaps that the awareness is more widespread.
The second dynamic that’s getting lost in the noise is that NOBODY should use today’s agents to make, say, Legal decisions for you. It doesn’t matter how sophisticated the output seems. If the success of your business is riding on it, a human must still be in the loop to catch the hallucinations and nuances the machine can’t. It speeds up your work, but it’s not capable of replacing it. (Side quest: Here’s Exhibit A for why AI isn’t replacing lawyers anytime soon, and why the Legal profession needs to be especially careful with AI.)
The third dynamic is that, as magic as build-your-own-app seems, building the product is just one piece of the puzzle, and it’s the fun part. There’s an entire operational (*cough* boring) system of resource management, security infrastructure, data connections and more that goes along with scaling it out. Not to mention, the expertise to know what you need. Even with agentic assistance, this is a lot to expect of your average business owner. They’ll still need ready-to-go platforms or other tech solutions that manage this for them, even if it means paying a premium.
Signs and portents
We avoided a true SaaSpocalypse this time around, which isn’t to say there won’t be one sometime in the future. Hidden among the falling stock prices are signs of bigger problems that could get really messy later on if we don’t address them now.
Cowork’s plugins represent the yeoman tasks we used to delegate to entry-level workers so they could learn the profession from the ground up. The judgment they cultivate by stepping through all the tactical paces is meant to get them ready for more strategic work as their career grows.
No one’s figured out a good replacement for this. Book learning can’t teach street learning, and intensive mentoring is a skillset senior professionals might not have the time or inclination to manage. Trouble is, this is a “we’ll deal with it later” problem for business and a “please deal with it now” problem for new grads. We’re stranding them to focus instead on implementing the technology. Eventually, the experienced seniors will retire, and we’ll wail and gnash our teeth that the juniors don’t have the acumen to fill their shoes.
Speaking of stranding folks, we’ve got an ongoing Jobpocalypse for knowledge workers, and particularly for knowledge workers who don’t fit the profile of white, straight, childless male. AI is taking the blame, but the truth is there’s a perfect storm of economic and societal factors at play and no clear end in sight. In the meantime, job seekers in survival mode are being forced to make incredibly tough choices that set aside what they want in order to pursue what they need. When the market turns, businesses will lose top talent. Frankly, they already have.
SaaS companies aren’t out of the woods either. To differentiate themselves from homegrown alternatives, they’ll have to un-enshittify their products and get really clear about their value prop and exactly why customers should stick with them. “Because it would be a pain to move to another tool” won’t cut it anymore if the alternative is cheaper, de-bloated, and tailor-made for what the customer wants.
Imagine that. To avoid the reaping, companies might have to put the customer first again.
All opinions here are my own. All text is my own, too, including the em dashes. I welcome constructive comments and discussion on LinkedIn and Bluesky.