Leverage AI buying intent

A lot of what we called a SaaS moat was really just the cost of building software. For two decades, a big part of software defensibility was that software was hard to build. If rebuilding the product was expensive, switching costs, workflow embed, and data lock-in held up pretty well. I think that frame is getting weaker. If AI agents do work that used to require more users, seat growth slows. If features are also easier to copy, per-seat SaaS gets structurally weaker with it. That does not mean defensibility disappears. It means it moves. My read is that the strongest moats now sit in three places: - money movement - compliance - infrastructure other companies depend on Those layers still take years to build because the constraint is not just code. It is licenses, audit history, bank relationships, approvals, and institutional trust. The screen I now use when looking at companies: what can a great engineer with Cursor and a weekend still not replicate? I wrote up the full argument here, including why I think this is where capital should be looking now: https://lnkd.in/eaJAaiuF