AI Pulse · Business Models & Winners

A pricing structure is making the rounds in AI-agency circles that looks a lot like what we already run at Enterprise DNA. The framework breaks into three pieces: a setup fee to cover architecture and build work, a monthly retainer that handles token costs and hosting, and a performance bonus tied to measurable business outcomes like attributed revenue.

The model showed up in a blog post that walks through the logic but does not name any agencies actually charging this way. That matters. It’s a thought exercise, not a field report. Still, the structure makes sense because it aligns cost with value at each stage. You pay once to build the system, you cover the real recurring expense of running it, and the agency earns more when the system drives results you can track.

We’ve been using a version of this inside the Omni Command Centre framework for a while now. The setup fee funds discovery, workflow mapping, and the initial agent build. The retainer keeps the system live and monitored without surprise invoices when usage spikes. The performance piece ties our work directly to outcomes like faster deal cycles or reduced support load, so both sides care about the same number going up.

The question is whether this becomes standard or stays niche. Most agencies still bill by the hour or quote fixed projects, which works fine until an AI system scales past the original scope. A three-part model forces everyone to think about ongoing cost and long-term value from the start. If more shops adopt it, expect clearer conversations about what AI infrastructure actually costs to run and what it should deliver in return. For now, it’s one blogger’s take and a handful of practitioners already doing it quietly.