Yelp prints your reply speed on your public page. Today our client's card reads: response time 10 minutes, response rate 100%.
That card is doing sales work. In our 150-business measurement across LA and NYC (August 2026), event businesses took a median of 2 hours to answer a Yelp lead. The customer messages three vendors and books the one that answers first, so the public card is not a vanity metric, it is the shortlist filter.
The workflow behind it runs on Zapier with Claude, handling about 600 enquiries a month (client-reported), for a Bay Area audiovisual event company:
1. Every Yelp lead is classified before anything replies: event type, date, size and whether it is something the business actually serves.
2. Qualification decides the path. Real, in-scope enquiries get a reply drafted in the business's own voice within minutes. Out-of-scope requests get a polite decline, so they stop eating the owner's attention between setups.
3. Qualified leads also fire a priority email to the owner, so the hottest conversations interrupt him and everything else waits politely in the log.
4. One follow-up goes out after 12 hours of silence, and a one-click takeover flag stops the agent on any thread the owner wants to run himself. The takeover is the feature: nobody trusts a bot that cannot be interrupted.
The honest part: the same public card still marks response quality as room to improve. Speed is solved, depth is the next iteration, and a system that shows you the next number to fix is exactly what you want running your inbox.
Yelp prints your reply speed on your public page. Today our client's card reads: response time 10 minutes, response rate 100%.
That card is doing sales work....