A new acquisition channel, validated on $396.70 of total spend.
The question was whether a conversational funnel could book qualified coaching calls at a cost the business could carry. The answer had to arrive before a media budget was committed, so the test was built to be cheap enough to be wrong.
214 conversations were run on $396.70 of total spend. The bot qualified on three things: what the person wanted to change, what they had already tried, and whether they could commit to a paid engagement. Anything that failed qualification was handed a resource and closed, so the cost per conversation stayed low enough that volume was affordable.
Every conversation was logged at the turn level, so the drop-off could be located rather than guessed at. The booking rate opened at zero, cleared 3.3% blended, and peaked at 4.5% once the qualification questions were reordered to put the commitment question last instead of second.
California's AI disclosure law changes what a conversational funnel is allowed to do and say. Rather than audit that by hand each quarter, an automated monitor watches the relevant regulatory sources and flags changes against the flows in production, so the disclosure copy is a maintained artifact instead of a memory.