The Domain Expertise Moat That Wasn't
February 19, 2026 by Asif Waliuddin
The Domain Expertise Moat That Wasn't
When Anthropic launched Claude for Healthcare, two things happened in the same week: OpenAI launched ChatGPT Health, and software stocks dropped.
Salesforce. Workday. The usual enterprise software suspects. Down.
The financial coverage described this as "competitive pressure." That framing is technically accurate and analytically useless. What actually happened is more interesting — and more structurally damaging to enterprise software valuations than "competitive pressure" implies.
The Hype
For fifteen years, vertical SaaS companies have sold a specific thesis to enterprise buyers and investors: domain expertise is a moat.
The argument went like this. Building software for healthcare is not like building general-purpose software. You need to understand clinical workflows. You need to understand HIPAA compliance in depth. You need relationships with health systems. You need to have seen the edge cases, the exception flows, the organizational politics of hospital procurement. This takes years. This is why you pay Veeva or Epic or Salesforce Health Cloud a premium — because they have absorbed that expertise and encoded it into their product.
This narrative powered billions in valuations. Investors accepted 10-15x revenue multiples for vertical SaaS on the premise that domain expertise was genuinely defensible.
The Reality
Anthropic shipped a HIPAA-compliant Claude for Healthcare with specialized agent skills for clinical trials, regulatory operations, and clinical documentation. They did not do this by acquiring a healthcare company. They did not do this by hiring hundreds of domain experts over a decade. They did it with a capable foundation model trained on medical literature and regulatory text, HIPAA compliance architecture, and a few months of focused development.
OpenAI launched ChatGPT Health the same week. Two companies simultaneously converged on the same vertical, independently, without the "years of relationship building" that vertical SaaS companies said was required.
Here is the mechanism that is being misread: AI does not need to replace Salesforce Health Cloud to damage its valuation. It just needs to make the "domain expertise" premium feel unjustified.
When an enterprise buyer looks at Salesforce Health Cloud and asks "what am I paying for that a Claude-powered workflow tool cannot do?" — and the answer is increasingly "switching costs and integrations, not intelligence" — the premium compresses. The valuation multiple compresses with it.
RBC Capital Markets put this precisely. They warned that vertical AI "challenges assumptions about domain expertise moats." That is not a product review. That is an investment thesis review. And when major banks start reviewing investment theses, valuations follow.
What This Means
The healthcare launch is not the story. Healthcare is just the clearest example because the regulatory complexity made vertical expertise seem most defensible. If Anthropic can compress that into a single product launch, every vertical is exposed.
Think through the stack:
Legal practice management: domain expertise in case law, compliance, billing. AI models have read every case and can apply compliance frameworks.
Construction project management: domain expertise in permitting, subcontractor coordination, change order management. Process-heavy, not intelligence-heavy.
Real estate transaction coordination: compliance-driven, documentation-heavy, relationship-dependent. Two of those three advantages are getting automated.
The vertical SaaS companies with durable positions are the ones whose value is genuinely in workflow integration — the data connections, the approval chains, the organizational processes that have been running through their system for five years. That is real switching cost. That is not being disrupted by a smarter model.
The companies that are in trouble are the ones whose pitch was fundamentally "we understand your industry." Because AI companies can now make that claim too, and they have better unit economics doing it.
The Bottom Line
The SaaS selloff was not irrational panic. It was the market updating on a specific hypothesis: that domain expertise, as a standalone moat, no longer justifies a premium.
The companies that will navigate this have genuine workflow lock-in. The ones that will struggle are the ones that have been selling expertise they can no longer claim exclusive access to.
RBC said it. The market priced it. The only question is which verticals discover this at product launch and which discover it at contract renewal.