> ## Content Index
> Fetch the complete content index at: https://futureinuse.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# AI agents in customer service: what the published L'Oréal and Crocs cases say about scope and reuse
- URL: https://futureinuse.ai/loreal-crocs-agentes-escopo-reuso/
- Published: 2026-10-14T12:21:59.000Z
- Updated: 2026-10-19T19:50:35.000Z
- Description: Two customer stories Salesforce published in September describe projects that started with a few tasks and grew from there. Both are cases released by the vendor itself, with numbers that have not been independently audited.
- Author: Rafael Cordeiro
- Tags: Agents at Work

In September 2026, Salesforce published two customer stories on its website about Agentforce, its AI agent platform, one on Crocs on September 10 and one on L'Oréal on September 11\. The information and numbers below come from those materials, written by the technology vendor.

## L'Oréal puts the agent inside the screen reps already used

According to Salesforce, L'Oréal's service reps spent time on routine tasks such as drafting emails, confirming the reason for contact and generating promo codes. Creating a single discount code meant clicking through about 15 screens, in a 10-minute process.

The agent was placed in Service Cloud, the tool the reps had been using for more than a decade. On each case, it summarizes the conversation, confirms what the consumer is asking for, generates the promo code from a text request and drafts the first email in each brand's tone. The rep reviews, edits and sends. A single deployment serves several brands, such as Lancôme, CeraVe and Kiehl's, with specific instructions for each one.

The published numbers show average handle time down 64%, from about 7 to about 2.5 minutes, a promo code now taking about 1 minute, and more than 300 reps using the agent every day, with adoption at 100% among eligible reps. The project began as a pilot in Canada with a single brand and now covers 25 brands in North America, according to the company.

## Crocs starts with five use cases and adds voice in seven weeks

In the Crocs case, Salesforce describes a customer-facing agent called Rivet, built in five months by Crocs, Salesforce engineers and the consultancy Rosetree Solutions. The launch focused on five use cases, namely product discovery, FAQs, order tracking, returns and conversational checkout.

The chosen architecture has specialized agents coordinated by an orchestrator. According to Rosetree, the idea was to make new use cases possible by adding a subagent. Crocs then took the agent to voice in seven weeks, reusing the actions and instructions built for the text-based service agent. Before every update, the company tests hundreds of conversation scenarios, with automated scoring against brand guidelines.

The material lists as an outcome an expected increase in add-to-cart rate, with no number attached. So far, the case describes build timelines, and measured business results are nowhere in the material.

## What the two cases have in common

Read together, the materials point to three design choices.

- **Narrow scope at the start.** One brand in one country at L'Oréal, five use cases at Crocs.
- **The agent where the work already happens.** The rep's screen in one case, the customer chat in the other.
- **Instructions and actions as reusable parts.** One deployment for many brands at L'Oréal, the same set carried from text to voice at Crocs.

Because these are stories published by the vendor, both materials leave out the measurement period, the baseline and consumer satisfaction. Even so, they help size a first project around a few repetitive, well-defined tasks and a record of the time spent beforehand, so the comparison later has a baseline of its own.

---

## Sources

- Salesforce, Customer Stories, "How Crocs built a retail superagent with Agentforce", September 10, 2026.
- Salesforce, Customer Stories, "L'Oréal handles consumer support cases 64% faster with Agentforce", September 11, 2026.

*Future in Use is an editorial initiative linked to OperOS AI.*