AI Will Reshape the Software Industry

Claude Code was launched (limited research preview) this February and then subsequently made generally available alongside Claude 4 a few weeks ago.

The AI coding breakthroughs will transform enterprise software over the next 2 years.  And when the next stronger version is released in the next 1-2 years, it will also have an effect on the public market valuations in ways that many may not expect but it will be a considerable market drop in the valuation of many companies.

There will be investors and industry pundits who will claim the death of software but but I only foresee the impact on certain software business models over the long term.

With AI coding tools advancing rapidly, a narrative is building that AI agents will render SaaS obsolete. I expect this to be partly right and mostly wrong. AI will reshape enterprise software, not replace it. Companies that own valuable data will emerge stronger.

Code Was Never the Constraint

  • AI reduces the marginal cost of code creation, similar to how the internet reduced distribution costs. But building a prototype is 1% of the work. Maintaining, securing, and scaling production systems is the other 99%.
  • AI-written code will not kill software companies. They will benefit by writing more code faster. The real threat is that every competitor now has this same capability.
  • What AI does not solve: GTM execution, customer acquisition, trust at scale, sales effectiveness, distribution, and customer experience.
  • A software company is not just code. Commercial execution is what creates a successful company, not coding speed.

AI Makes Strong Companies Stronger, Weak Companies Weaker

  • The moat of durable software companies was never the software. It was proprietary data, workflow integration, enterprise lock-in, network effects, trust, compliance, and deep customer relationships.
  • AI amplifies these advantages for strong companies. For weak companies whose only asset was a UI over a standardized process, AI makes them expendable.
  • Most at risk: Workflows where automation is easy and third parties can replicate the functionality. Tier 1 support, invoice processing, time-entry approvals, basic task management.
  • Most insulated: Workflows with deep domain judgment, regulatory complexity, and proprietary data. Construction cost accounting, clinical trials, insurance claims adjudication.
  • The dividing line: if you own defensible data and embedded business rules, AI strengthens your position. If you own the UI to a standardized process, AI threatens it.

SaaS Companies Sell Business Processes, Not Code

  • SaaS companies are selling Business Processes as a Service. The value is operational reliability, security, and integration depth that enterprises cannot replicate internally.
  • Salesforce, ServiceNow, and Workday have survived despite cheaper alternatives (Sugar CRM, Zoho) existing for 15+ years. The moat is bundled operational capability, not technology.
  • When you acquire a vertical software company, you are buying the “stored process”: institutionalized workflow knowledge that is expensive to recreate and risky to abandon. AI coding breakthroughs do not touch that.

The Enterprise Trust Problem

  • Large enterprises move slowly on critical workflows. A CRM with 10 years of institutional knowledge will not be replaced overnight by an agent layer, even a technically superior one.
  • Incumbents will not stand still. They will build their own agent layers on top of their data moats.

Where Disruption Will Land

  • Systems of record (ERP, CRM, HRIS) remain essential but get pushed into the background as AI orchestration builds on top
  • Systems of engagement (UIs, dashboards) face compression as AI transforms how humans interact with software
  • Pricing models shift from seat-based to consumption-based as agents replace human users
  • AI players may capture the value layer on top of systems of record, shifting where economics accrue
  • This will not happen overnight, especially in regulated industries. But the broader pattern will be market expansion, not contraction. There will be more software than ever.

Bottom Line for PE Investors

Software is not dead. But retention, terminal value, and multiples will be lower because AI disruption potential lowers predictability of future cash flows. This only changes if companies demonstrate how AI accelerates their core business. Focus diligence on:

  • The operational complexity being solved, not just the software features
  • Switching costs tied to business process integration, not just data lock-in
  • How AI capabilities expand the problem set the company can address
  • Whether the company owns binding data and state, or just a user interface
  • Whether the business can monetize agent-driven usage without cannibalizing existing economics