The SaaS Doom Bubble

One notable feature of the AI bubble is how disconnected the markets are from day-to-day corporate reality. Random tech stocks keep on being hit because of the latest AI advancements. Every week sees a new model or tech demo from a leading AI lab. The highlighted capability often competes with a range of existing SaaS solutions, which it is claimed to replace. Investors panic. Dumping the shares of unrelated SaaS companies, including vendors such as Salesforce which are used internally by the leading AI labs.

AI has resulted in a new wave of tech startups, many of which compete directly with the leading SaaS companies. The likes of Clay and 11x offer to replace traditional sales automation tools. While some existing vendors such as Qualified have gained new relevance from the technology. Yet, forget the legacy players. Investors are acting as if all these firms will also be replaced by vibe coded in-house solutions. More likely, they will be absorbed by incumbent vendors such as Salesforce. Following on from the CRM giant’s purchase of Qualified at the end of last year.

The Return of In-House Software?

What AI does is re-open the debate about buying versus building software. This was once a live argument played out through flame wars about the merits of Access databases. The rise of cheap SaaS software eliminated the cost advantage of in-house apps, especially for non-tech firms. While vibe coding technology is vastly overhyped, it has once again transformed software development workflows and revolutionised the economics of in-house software. As a result, it is much easier to build simple apps. The trouble is that a lot of enterprise software is far from simple

Most enterprise CRM systems are complex beasts with sprawling data maintenance processes and vast territory management capabilities. Security and compliance oversight is mandated at every stage. These tools touch every part of the business. Day-to-day operations are powered by the default logic native to the platform. Few companies use every feature of Salesforce but most enterprise customers use enough of the platform to make migrating a multi-year project.

The Risk for Add-On Vendors

Businesses standardise on Salesforce because of the perceived flexibility, the enterprise-grade security and the easy availability of business-critical support. Vibe coding only neutralises the first of these three advantages. That's even before the vast software and services ecosystem built around the platform is considered. For all its complexity, Salesforce’s native capabilities often aren't sophisticated enough. An entire industry of add-on vendors has built up because Salesforce’s standard routing and attribution workflows simply don't scale to use cases in some industries.

Some of these add-on vendors will be threatened by the rise of vibe coded apps. If your app offers one feature built on top of another vendor’s platform then it may be possible for it to be replicated by a moderately technical user with generative AI. It may already be possible to replicate your add-on through bespoke customisation or clever workarounds. For that reason, the life of an add-on developer has long been precarious anyway. If your add-on is popular, you're always at risk of being made redundant by the platform owner’s next release. AI merely compounds that risk.

A Mature SaaS Market

The real threat to the SaaS industry is not AI. Where it makes sense, vendors have already incorporated AI into their apps. Home-brewed apps are not a threat to SaaS either. They've always existed, and always will. Few businesses will use vibe coded apps for core business platforms because of the compliance risks and the difficulty of maintenance. The real threat is market maturity. SaaS was popularised by three leading technology trends of the 2010s: digital transformation, cloud migration, and big data. Technology teams adopted vast tech stacks in response. Yet, that migration process was slowing long before generative AI became the hot new technology. Tech stack consolidation was the buzz phrase of the 2020s.

Tech firms have invested enormous sums into generative AI, precisely because they have few other sources of growth. Product bundling and suite building emerged as a response to this problem. Yet, that strategy merely cannibalises other vendors. It doesn't grow the overall market. Tech bosses are desperate for AI to solve their growth problems. They're probably going to be disappointed. Prior to the launch of Agentforce, Salesforce was under immense pressure to slow down product development and return cash to shareholders. Recent share price corrections are the first sign of a return to pre-bubble market trends.