All cases

Customer Service

Lacoste's Ask René agent resolves 60% of chats and briefs advisors on the rest

60%

“Four days after launch on the US market, early indicators are encouraging: Consumer satisfaction: above 4/5, exceeding the target set. Escalation rate: on track for -30% compared to the previous chatbot. Handling time for complex cases: under 8 minutes on the human advisor side. [...] Production metrics validate the operating model: 60% autonomous resolution and an escalation rate capped at 30%.”

Converteo, Apr 2026; Converteo, May 2026, 27 Apr 2026

Prompt

You are the customer service assistant for our [tennis / padel / golf] brand's online shop. My input: [return and warranty policy], [shipping and order-tracking rules], [top 20 customer questions], [brand tone: words to use and avoid], [when to hand over to a person]. Give me: 1) answers to the top questions in our tone; 2) a checklist to qualify a product-defect return (photos, order number, usage); 3) the summary note you leave for the advisor when you hand over; 4) the signs of an upset customer that mean a person takes over now.
Who
Lacoste, the fashion brand founded by tennis champion René Lacoste, on its US website. The agent, Ask René, was built with the consultancy Converteo and Google Cloud and went live on lacoste.com in the US in early 2026.
Problem
Lacoste's previous chatbot on the US site could only handle after-sales and could not help with product discovery or purchases, far from the welcome of a Lacoste boutique. It also had to be switched off by March 2026, so a replacement had to be live within months.
Agent
A central orchestrator routes each request to an expert agent: a knowledge agent for FAQs and return policy, an order-tracking agent connected to Manhattan Associates, and a product-quality agent that qualifies defect returns and prepares the case. If the customer gets irritated, the conversation goes straight to a human advisor in Salesforce Service Cloud with the full history.
Result
Four days after launch: satisfaction above 4/5, escalation on track for -30% vs the old chatbot, complex cases handled in under 8 minutes by advisors. A few weeks later, Converteo reports 60% autonomous resolution with escalation capped at 30%. Lacoste plans Canada, then Europe and Latin America.

We'll build it for you

Book a call
What we can do for your club or brand

Racket and golf brands can let an agent answer routine questions and prepare return cases so their service team spends its time on the customers who need a person. The Court can turn that into a sourced public case and the content that shows customers how the brand looks after them.

Get it implemented

Source:

Converteo, Apr 2026; Converteo, May 2026, 27 Apr 2026 · vendor case study

Is Google Gemini Enterprise for CX (orchestrator + expert agents) your product? Get it featured

What they did

Lacoste's teams and Converteo built Ask René in about four months on Google's Gemini Enterprise for CX, working inside Lacoste's digital factory. They connected it to Salesforce Commerce Cloud, Salesforce Service Cloud and Manhattan Associates, trained it on Lacoste's brand voice and escalation wording, and launched it on the US site before the old chatbot was switched off in March 2026.

How it went

Converteo reports satisfaction above 4/5 from the first days, escalation on track for -30% and complex cases handled in under 8 minutes on the advisor side, then 60% autonomous resolution with escalation capped at 30%. Advisors receive the whole conversation and photos with no re-typing, and upset customers go straight to a person. Limits: both sources are the integrator's blog; Lacoste has not published its own figures, and the metrics come from different moments after launch.

Want this for your club or brand? The Court builds it in 2 weeks.

Book a call

The 5 best cases of the week, every Sunday

Real tennis and padel AI agent use cases, each with its source and the prompt to try it. Agents that augment the team, not a replacement. One email, every Sunday morning.

Free. Unsubscribe in one click.