Users rarely seek out separate AI functionality. Adoption becomes natural when AI assistance appears contextually within existing workflows, addressing friction points directly where the user is already working. This embedded approach is far more effective than adding AI as a separate, layered-on tool.

Traditional SaaS switching costs were based on painful data migrations, which LLMs may now automate. The new moat for AI companies is creating deep, customized integrations into a customer's unique operational workflows. This is achieved through long, hands-on pilot periods that make the AI solution indispensable and hard to replace.

Don't just sprinkle AI features onto your existing product ('AI at the edge'). Transformative companies rethink workflows and shrink their old codebase, making the LLM a core part of the solution. This is about re-architecting the solution from the ground up, not just enhancing it.

The SaaS-era advice to "do one thing well" is outdated and risky in the current AI climate. The best defense against rapid displacement by competitors or platform shifts is to build a multi-product bundle. This strategy creates a wider surface area within a customer's workflow, increasing stickiness and defensibility.

The most defensible AI companies don't just have superior models; they embed themselves deeply into customer workflows. The primary barrier to adoption is change management, so overcoming that hurdle creates a durable competitive advantage that is difficult to displace.

An impressive AI capability, like a multi-language voice agent, is a differentiator that can be copied. Lasting defensibility is achieved not by the AI feature itself, but by embedding it within an end-to-end workflow that becomes the system of record for the user.

Software's main competitive advantage isn't code, but its deep integration into customer data and workflows, creating high switching costs. AI threatens this moat by automating those integrated tasks, reducing customer stickiness and pricing power.

To get mainstream users to adopt AI, you can't ask them to learn a new workflow. The key is to integrate AI capabilities directly into the tools and processes they already use. AI should augment their current job, not feel like a separate, new task they have to perform.

Linear doesn't try to build a better general-purpose coding agent than Google or OpenAI. Instead, its strategic advantage is sitting 'upstream' where work originates. By integrating agents into the initial bug report or feature request, they can automate the entire workflow, a defensible moat.