If I Wanted to Build an AI SaaS in 2026, I’d Do This

Key Concepts

• Surface to Software Ladder: A three-phase framework for building B2B SaaS by starting with manual, high-touch services and evolving into automated software.
• Breakage: A specific business problem characterized by being painful, repetitive, costly, and currently unfixed.
• Product-Market Fit (PMF): The state where customers pay for a solution and continue to use it because it provides tangible value.
• Done-For-Them (DFT): The initial phase of manual service delivery to validate a problem.
• Done-With-Them (DWT): The second phase involving semi-automated solutions for a small group of similar clients.
• Done-By-Them (DBT): The final phase of a fully self-service, scalable SaaS product.
• Turn-It-Off Test: A validation metric where the product is so essential that its failure triggers an immediate, urgent response from the client.

---

1. The "Surface to Software" Ladder

The author argues that most developers fail because they build in isolation for months without validating market demand. The "Surface to Software" ladder reverses this by prioritizing revenue and proof over code.

Phase 1: Done-For-Them (The "Surface" Phase)

• Objective: Identify "breakage" within a company and solve it manually.
• Methodology: Avoid building generic tools (e.g., task managers). Instead, use your warm network to find insiders who can point to messy, unsexy, or inefficient internal processes.
• Discovery: Position yourself as "curious" rather than a salesperson. Ask questions to uncover problems that are painful, repetitive, and costly.
• Validation: Propose a Proof of Concept (POC) with a clear price tag. If a client is unwilling to pay for the solution, the problem is not significant enough to warrant a SaaS product.

Phase 2: Done-With-Them

• Objective: Acquire 3–5 near-identical customers to refine the product skeleton.
• Methodology: Rebuild the solution for similar companies in the same niche (e.g., e-commerce stores using a specific ticketing system).
• Pricing: Shift from flat build fees to value-based pricing (e.g., charging per ticket processed). This aligns your revenue with the value the client receives.
• Outcome: Collect testimonials and data to prove the model works beyond a single client.

Phase 3: Done-By-Them

• Objective: Transition to a fully automated, self-service SaaS.
• Methodology: Focus on engineering the "surrounding" complexity: onboarding flows, multi-tenant support, password resets, and magic links.
• Key Metric: Time-to-Value (TTV). Optimize for how quickly a new user can achieve a "quick win" after signing up.
• Distribution: Move beyond paid ads. Focus on built-in distribution mechanisms like network effects or referral incentives (e.g., lifetime commission models).

---

2. Important Considerations

• The "Turn-It-Off" Test: If you disable your software and the client doesn't immediately call to complain, the product is not essential. A successful product should be so integrated into their workflow that its absence causes a 4x increase in their manual workload.
• Compliance: As you scale, security becomes a barrier to entry. The author highlights the importance of frameworks like SOC 2, ISO 27001, and AI governance (e.g., NIST AI RMF) to close deals with larger enterprises.
• Bootstrapping: By following this ladder, you retain 100% equity and avoid the need for external funding, as the business is self-funded by early service revenue.

---

3. Synthesis and Conclusion

The core argument is that code is the easiest part of the business. The true challenge is finding a customer, getting them to pay, and ensuring they stay. By starting as a freelancer or agency (Phase 1), you get paid to conduct market research. By moving to a semi-automated model (Phase 2), you validate the product's scalability. Only after these steps are complete should you invest in the heavy engineering required for a fully automated SaaS (Phase 3). This approach minimizes risk, ensures product-market fit, and allows for a sustainable, bootstrapped business model.