AI Agents Have Changed How Tasks Are Approach...

Why We Still Buy Software in the Age of AI

If anyone can make anything with AI, why do we still pay for software?

Why We Still Buy Software in the Age of AI

If anyone can make anything with AI, why do we still pay for software?

If you are, or have been, in the SaaS world, you probably remember the chill from that day in February when SaaS companies were hit hard by the "SaaSpocalypse", that period when the Market was betting against SaaS companies because AI could replace their products so easily. The premise was hard to ignore. LinkedIn posts were popping up daily on this or that expert rebuilding the platforms that were ServiceNow, Salesforce, Workday, and Palantir, all within a weekend. The cost? Several thousand dollars worth of tokens. The question then became, "Why would you buy expensive licenses when you could build your own replacement?"

That argument persists today. More and more specialists, often company employees who had been let go due to "restructuring" built similar platforms as their prior companies, often at the cost of several thousand dollars worth of tokens. Their goal? Get hired again by showing what they can do with tokens. The Media's take? Modern SaaS is on thin ice, as companies will start building their own replacements.

Thanks for reading Robbclan Substack! Subscribe for free to receive new posts and support my work.

Commercial vs. Homemade: The History Lesson

The best analogy I can make is the shift in the United States from Colonial homemade products to post-Civil War, commercially produced items. During the Colonial period, most of the United States was comprised of settlers, frontiersmen that had to make do by themselves. They made their own clothes, preserved their own food, and even made their own soap, all by hand and with what they had available. The results were, well, sufficient for the need, or at least were sufficient enough for survival. Just not convenient.

As communities grew, you would get those settlers who could specialize: Perhaps the Perkins family could make soap better than anyone else, while the Davies family were excellent spinners and weavers. Their expertise meant results were significantly better than homemade products could be given the available resources, and more affluent settlers would opt to purchase products, like soap, from those who were more skilled at making those products. Wealth allowed for specialization, and specialization allowed for better product development. Anyone could make soap, but only the Perkins family could make it smell so fresh.

Industrialization meant that those specialty families could start automating a lot of their processes, making more uniform results at a higher quality. The expertise they built in knowing the product development inside and out became invaluable insights while orchestrating the mass-production of that same product. By the end of the Civil War, as we moved into the Roaring '90's, then the Roaring 20's, specialization was mechanized. Sure, you could still make your soap at home from the ashes of your fireplace, but mass-produced soap smelled better.

Homebrew vs. SaaS Software: The Parallel

Computers and software went through the same thing. For years, even from the days of punch cards, commercial software was homebrewed within the company, built for the specific tasks the company made with their own workflows and processes, good or bad. Throughout the 70's, Mainframes were the primary source of software, with precious few terminals available for those few experts who would know the Fortran or COBOL necessary to code and maintain the system.

In the 90's, personal computers and networked systems with servers running in racks were the primary setup, with custom software facilitating the communication. By now, Office suite documents were sold for their convenience: You had set processes and interface tools that made building content easier, then sharing it through email (I still cringe at those days) or through shared storage space on the company Server. All other office or company functions? They were generally custom-built for the company.

Then, on-premise software systems were released to manage business, made by companies who were building custom business process platforms, because they had the experience. Companies would purchase this software because it was more reliable, more polished, and provided a uniform workflow for companies to follow.

The next iteration came with SaaS platforms, which moved that expertise and workflow standards off-site (meaning less maintenance for teams) and into the cloud for more reliability and less down-time. The expertise, both in the workflows and in coding, were necessary, they were just off-loaded to specialists, much like the settlers buying soap from the Perkins family.

Age of AI: Expertise Shifts, Specialization Shines

So let's bring this full circle. In February, it was predicted that AI was going to make SaaS unnecessary, because experts in their field were building replacements over the weekend. Yet, ServiceNow, Salesforce, Workday, and many others are still here. Why? Because, while AI lowers the expertise in coding, it doesn't replace the knowledge and judgement that comes from specialists.

Yet, companies have let many of their senior employees go through restructuring, others left because they saw opportunity. Now, you are getting several new AI-powered and Agentic-friendly platforms developed and released from companies of 1-5 employees, providing similar services with new Agentic capabilities. The barrier to development is dropping, and now expertise and specialization are shining.

Where We Are Heading

The future of work is shifting. Work will be conducted within an Agent prompt. Not might be, it will be. I say this confidently, because it makes the most sense. We've been slowly pushing actual work to smaller and smaller devices to make it easier to conduct work while on the move. Agents, accessible from an internet-connected device, have the capability to do the work we ask it do to, and the convenience to do it from a prompt. On a phone. While skiing. The future of work will reflect that, as experts who have specialized in workflows of various types provide their agentic solutions. The Agent uses those solutions for you (so you don't have to log into 15 websites every morning), and you orchestrate your work based on your expertise.

The shift here? SaaS will always be there for large corporate entities. They have the contracts, the name recognition, and often the capability to pivot to the future if they act fast enough (not all legacy software does). The smaller outfits, developed by those experts that have left those large SaaS companies? They will start with smaller markets, but because they are more agile in their approach, will quickly adjust and build to the needs of their customers faster. AI makes that simple, often with hourly updates and patches. That agility, provided they can scale, will be the real superpower.

The Better Soap

Look, the cycle we are in isn't new. It seems like it is, because it's moving fast and we are in the middle, but using the Historical perspective, it's the same shift that came with Colonial soap-making to industrial soap products: Specialization and expertise are the differentiators here. The new market will be more flexible, move faster, and will come from experts who have been in the space a long time and see the opportunity to move fast that traditional SaaS can't or won't pivot toward.

Sometimes, that means starting with your homemade "soap" and making it better than everyone else's.

Thanks for reading Robbclan Substack! Subscribe for free to receive new posts and support my work.

Share

Subscribe for free to receive new posts and support my work.