The AI Differentiation Gap: Why Build-Your-Own Beats Vendor Licensing in Commercial AI Strategy

If your AI strategy depends entirely on third-party vendors, you may be building a commodity position rather than a competitive one. That's the argument John Swigart, co-founder and CEO of Pie Insurance, is making for where commercial lines AI investment goes next.

The Core Argument

Swigart's thesis is direct: "If you only use software and tools provided by third-party companies, you will only ever look the same as the other customers of those companies." AI has lowered the cost of building proprietary software enough that a single AI-fluent developer can now build what previously required a full engineering team.

Key Takeaways

• Build vs buy is the central AI strategy question. Bespoke models require massive data and capital, creating real barriers for smaller organizations. But operational AI adoption through experimentation is accessible to most.
• Agentic AI is already delivering. Deloitte-cited research found early agentic AI deployments in insurance produced underwriting efficiency gains up to 36% and claims cycle-time reductions near 40%.
• Four competitive fronts are emerging: proprietary in-house software, bespoke models backed by licensed third-party data, broad operational AI adoption through experimentation, and data pipelines precise enough to catch classification errors in real time.

Swigart's advice on where to start: "The key is to just start doing it, experimenting, finding use cases." Top-down planning is not the entry point. Experimentation is.

This build-vs-buy argument extends well beyond insurance. Any industry where workflow and data are core to competitive positioning faces the same fork: build a proprietary edge or rent parity from a vendor.

Read the full article on Insurance Business