Last week I ran into an old pal at a party who asked me what I thought of the “SaaSpocalypse” – the AI-driven correction that hit nearly all software stocks over the past few quarters. My friend has bought and sold software companies for most of his career and is a sophisticated player in private equity markets. Like everyone in finance and tech, he reads just about everything he can about AI’s impact on our economy, and in particular on formerly predictable sectors like software. “Software companies are cheap right now,” he said, before adding that he wasn’t sure it made sense to buy them, because everything he was reading was saying the software party is over.
It might have been the cocktails talking, but before I had a chance to think, I said “AI is just bad software. It’ll get better, once companies figure out what to actually do with it.”
Anyone saying that the software industry is over is failing to pay attention to what software is actually good for: Driving products and services that enterprises and consumers find valuable. The idea that somehow AI is going to replace all that with a chatbot interface – “build me a new supply chain system, Claude” – strikes me as ludicrous. Today I’m going to think out loud about why.
First of all, AI is software. It’s driven by code, algorithms, data, and compute. At the moment, AI software is not reliable: it handles long-term memory and complicated tasks poorly, and it’s a nightmare to integrate into an enterprise’s current digital systems. But that’s changing, rapidly. Many are predicting that in the near future, AI software will replace Salesforce, Atlassian, Adobe, and countless others. How exactly that replacement will happen is often left to imaginations fueled by excessive optimism and a misguided understanding of how corporations actually adapt to new technologies.
As every corporate leader I’ve talked with has already learned, AI isn’t really a product. It’s a capability, one that must be integrated into a company’s workflows, often changing those workflows as new possibilities are discovered.
Powerful new technologies have always driven a re-factoring of how corporations run their businesses. Those shifts usually meant leaving old processes, investments, and job titles behind. First was digitization itself – moving from paper-based to digital accounting, payroll, and inventory management meant abandoning not only rooms full of filing cabinets and typewriters, it meant retraining and/or laying off millions of office workers. Firms like IBM went from selling office equipment to selling mainframe timesharing. If you had bailed on IBM as an investment in the 1950s because of a mainframe-driven “typewriterpocalypse,” well, I guess you missed out.
The shift to the personal computer was equally impactful, as was the rise of the Internet, the migration to the cloud, and the introduction of the smart phone. Yes, thousands of companies went by the wayside during the dot com crash, but if you held onto Amazon, eBay, and Adobe through that crash, you’d be doing quite nicely, thank you very much.
But back to my friend’s original question: Is the software industry screwed? Absolutely not. I can imagine no scenario where businesses and consumers abandon using software tools to get things done. What I can imagine is a totally re-factored software sector that integrates intelligence throughout its product lines. The faster companies figure out how to do that in ways that create value for their customers, the more those companies will prosper. The ones that fight rearguard actions to protect a shrinking customer base will lose share and ultimately fail.
In other words, same as it ever was.
This past week the market began to come to its senses as a number of software leaders, their stock prices battered by months of “SaaSpocalypse” rumblings, reported record earnings buoyed by successful pivots into AI-enabled opportunities. Those stocks – Cloudflare, Atlassian, and yes, even Salesforce – are harbingers of the future. Companies are indeed figuring out how to actually use AI, and they’re employing software to get the job done.
—