Eight commits repositioned the page from education surface to qualification surface — booking flow consolidated, pricing plain-language, and the LLM Council Flywheel pipeline ungated.
Eight commits across the aicoachellavalley homepage repo. V1 of /get-agent-ready/ read like a brochure — it explained what the program was and left the reader to figure out next steps. V2 operates as a qualification surface: a visitor arrives, reads the criteria, sees the pricing in plain language, and books directly without leaving the page. The booking flow is no longer a secondary action buried below the fold. It's the destination.
The LLM Council Flywheel — the pipeline that routes qualified operators into the AICV network, connects them to peer cohorts, and surfaces their use cases as regional intelligence — moved from gated to ungated. It was previously described only in the intake form. It's now documented on the page itself, so a founder reading it understands what they're qualifying for, not just that they're applying.
The visualizer ran 15 iterations on the qualification criteria block before the layout locked. The core problem was hierarchy: the page had to communicate program rigor (this is not a workshop, this is a working engagement) without reading as bureaucratic. The solution was a two-column criteria grid with explicit pass/fail framing — a founder can self-select in or out in under thirty seconds. That framing didn't exist in V1.
Pricing went from vague directional language to a plain-language rate card with context: what the engagement includes, what the output looks like, and what the comparison is to hiring equivalent capacity as staff. The context matters. Without it, the number reads as a line item. With it, it reads as a build decision.
Two architectural decisions won't be revisited. First: the page does not explain AI. There are fifteen other pages on aicoachellavalley.com that explain AI to the Coachella Valley. /get-agent-ready/ assumes the reader already believes. The job of the page is to qualify and convert, not to educate. Every V1 element that explained AI context was removed in V2. Second: the booking CTA does not ask for a demo. It books a working session. The framing shift is intentional — demo implies evaluation of the vendor, working session implies evaluation of the operator's readiness. That's the right frame for this program.
The qualification-surface pattern is reusable. Any AICV program page that has both a criteria layer and a conversion action should be built this way: criteria visible above the fold, booking inline, explanation subordinated to qualification. The LLM Council Flywheel documentation on the page also sets a precedent — program infrastructure should be legible to the people being asked to enter it. That was underweighted in V1 and is now a design constraint going forward.