Will AI Kill Vertical SaaS?
How to prepare for the uncertain technical advances of AI.
This newsletter is 100% written by me. No ghostwriters or GPTs.
Happy Sunday and welcome to Investing in AI. I’m Rob May, CEO at Nova. We make a brand safety tool for generative AI. I’m also an active angel investor in the AI space. If you are working on something cool I’d love to take a look.
Recently I’ve seen a lot of companies who are using AI, and often GPT-4 in particular, to write code. It’s an exciting trend, but it has me thinking… if over the next 5 years the cost to create software drops 90%, what are the repercussions?
I see three things that could happen in this future world.
Vertical SaaS becomes a bad place to invest. Part of the reason SaaS is such a great business model is that switching costs are high. The migration off a typical marketing automation system usually has a 2-3 year payback, so companies don’t move for financial reasons. But if writing the migration code, and executing on the migration takes just a couple of days with AI-based development tools, will companies switch more? Will churn rates go up?
Build vs buy decisions skew towards build. Today, it rarely makes sense to build an application internally rather than buy a SaaS tool that meets most of your needs. It takes too long and costs too much and the maintenance costs are too high. What if AI-based coding tools in the next couple of years can help you build your own CRM in days? Would you do it to get exactly what you want? What does that mean for software companies in general if building becomes cheaper than buying? Will AI tools be able to do software maintenance work as well as they do the initial build?
Non-venture backable software companies could explode. If the price of software drops dramatically, that means we will get more software. Most likely that means more fragmented software markets, and that means perhaps we see a lot of $15M revenue software companies with 5 employees and 30% net margins that can’t grow at venture scale.
Many factors beyond the price of software creation will impact this future. Customer acquisition costs still play a role, and maybe at scale those come down so much that big software companies still dominate. And these observations really only apply to workflow applications. Marketplaces and networks will still be valuable because they are more than software.
Maybe the prompt precision required to create custom software creates so much complexity and overhead - that the benefits of rapid low-cost software creation aren’t fully realized because of the requirements→prompts bottleneck.
As an early stage investor in AI companies, I never try to predict the future. I just try to understand the possible futures that could be created, and try to get upside exposure to the best ones. Or maybe in the case of the ideas above, avoid exposure to areas likely to crash.
If you’ve thought about this, I’d love to hear your feedback. Thanks for reading.
Insightful article.
Memo to myself: https://share.glasp.co/kei/?p=d0mjZmLwkr8bmI7pJr5H
Don’t for get, a business value proposition is more than the software product. I don’t think companies will want to internalize the product management, roadmap prioritization, support, maintenance and operations of managing dozens and dozens of custom apps. It won’t be their core competency.
I do think companies could invest in custom apps that are very, very particular to their domain / industry.
Perhaps more interesting would be for traditional cloud software vendors to offer higher degrees of customization through dynamic AI-defined modules. As the software provider your cost to manage a more customized solution could go down dramatically.