Skip to main content
Large language models capable of generating software on demand will shift where value lives in the technology industry — away from pre-packaged software features and towards the data, services, and APIs behind the scenes.

​
What this piece is

This paper argues that large language models capable of generating software on demand will shift where value lives in the technology industry. Instead of value residing in pre-packaged software features, it will move to the data and services behind the scenes — accessible via APIs. The paper introduces the concept of Software-as-a-Prompt (SaaP) and examines what this shift means for startups. Originally published on Medium, August 2025.

​
What is the central argument?

If AI can generate functionality on demand, the code itself becomes a commodity. What remains scarce and valuable is the data feeding the AI and the real-world services behind the code. Startups that only offer a front-end with no unique backend are at risk of being bypassed. The future belongs to those who control data, infrastructure, and API access — while AI handles the functionality layer.

​
What are the key concepts?

Software-as-a-Prompt (SaaP). Instead of subscribing to software someone else built, you describe what you need and the software builds itself. You no longer adapt your workflow to the software — the software adapts to you. The API-centric world. In a future where AI assistants handle user interaction, startups thrive by supplying the intelligence and capabilities behind those assistants. Being a great API provider — reliable, well-documented, fast — becomes a competitive edge. What remains defensible. Proprietary datasets, network effects, enterprise-grade reliability, and unique integrations with the physical world. These assets cannot be generated by a prompt. Startup survival strategies. Embrace AI-first thinking. Offer your product as an API. Leverage proprietary data and feedback loops. Focus on network effects and ecosystem. Be the control plane for AI-generated software — handling security, compliance, versioning, and reliability.

​
How does this connect to other wiki topics?

This paper connects to the startup operating model thinking in Operating models and execution and the commercial readiness work in Fintech GTM. The API-first strategy aligns with the platform infrastructure approach used in Open banking and identity. The startup survival strategies map directly to the lean validation approach described in Leveraging GenAI and Distributed Teams.