Yelp

Objective

Rebuild the journey around answers and actions rather than search alone. Yelp's commercial value rests on the quality of the matches it sells to local businesses: a search that does not lead to a quote request earns nothing. The assistant captures intent while it is still vague, shapes it into a qualified request, and thereby feeds the flow of paid leads, in a pressured local advertising market where growth comes from other revenue rather than from legacy advertising.

The deployment

Yelp a fait de son assistant conversationnel le nouveau point d'entree de son moteur de mise en relation. Yelp Assistant dialogue avec le consommateur pour cerner un besoin de service local, puis met ce besoin en forme comme une demande de devis, le flux Request-a-Quote, adressee aux entreprises pertinentes. Ces entreprises sont les clientes payantes de Yelp : chaque demande qualifiee est un lead qui alimente le modele publicitaire. Au premier trimestre 2026, l'assistant a ete etendu a toutes les categories, et il represente deja environ 15% des Request-a-Quote du trimestre, contre 5% un an plus tot d'apres l'echange en earnings call. La direction relie explicitement l'assistant a la bascule strategique de l'entreprise, qu'elle decrit comme une reconception autour des reponses et des actions. Cote resultats, la croissance ne vient plus de la publicite locale, restee quasi stable (chiffre d'affaires net de 361 millions de dollars, plus 1% sur un an), mais de l'other revenue, qui progresse de 75% sur un an pour atteindre un record de 29 millions de dollars, portee par Yelp Assistant, la solution de reponse aux leads Hatch, l'offre Yelp Host et la cession de donnees. Yelp fixe une cible de 250 millions de dollars de run rate annuel d'other revenue d'ici fin 2028. Ce que le cas ne dit pas : l'effet net sur le revenu publicitaire par entreprise, ni le taux de transformation d'un devis genere par l'assistant en client servi.

Results Proof A

The case rests on Yelp's first quarter 2026 results, filed with the SEC (Form 8-K, Exhibit 99.1) and discussed on the earnings call of May 7, 2026. The primary release gives other revenue growth (up 75%, $29 million), the extension of Yelp Assistant to all categories, and the target of $250 million by end of 2028. The 15% share of Request-a-Quote carried by the assistant, versus 5% a year earlier, comes from the shareholder letter and the earnings call exchange. Honest caveat: the other revenue line aggregates several components (Yelp Assistant, Hatch, Yelp Host, data licensing), so the revenue specific to the assistant is not isolated; the cleanest attribution is the share of Request-a-Quote, a usage metric.

The internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.

1. 1
   Expressing the need client

   The consumer describes a local service need to the assistant, in natural language.

2. 2
   Intent qualification AI

   Yelp Assistant dialogues to pin down the project (nature, context) and shapes it.

3. 3
   Quote request generation AI

   The assistant turns the qualified need into a Request-a-Quote sent to the relevant businesses.

4. 4
   Reception and response on the business side AI

   Advertiser businesses receive the lead and respond, with a response that can be automated via Hatch.

5. 5
   Matching client

   The consumer compares responses and picks a provider, which closes the conversion loop.