The Italian automotive dealership group Autotorino has entrusted the handling of a portion of its calls to a voice agent built on ElevenLabs’ ElevenAgents platform. This agent answers calls, qualifies requests, and creates business contacts, known as ‘leads’. The group receives around 100,000 calls per month across its dealership network, and its team cannot answer them all.
Founded in the 1960s, Autotorino claims a long tradition of innovation. The group was the first Italian company to feature in the ranking of the top 50 European dealer groups established by ICDP, the first in Italy to import Japanese SUVs, and the first to present competing brands in the same showroom. Before the voice agent, its system relied on Zapier, coupled with separate transcription and language processing tools, to transform a recorded message into a prospect in Salesforce. The group wanted callers to be able to converse with an agent, while maintaining the goal of qualifying each request.
A choice based on three criteria
Autotorino compared several providers before selecting ElevenLabs. The group evaluated each solution based on three criteria: latency, meaning response time, the human-like nature of the interaction, and the system’s ability to understand what the caller is saying without transferring data between multiple tools. ElevenAgents was ultimately chosen as the platform.
Now, the customer service agent answers calls around the clock. Previously, calls received outside opening hours often went unanswered, leaving customers without any service. Michele Spalla, Head of Project Management and Innovation at Autotorino, summarizes: ‘Thanks to ElevenLabs, we are able to respond 24/7, providing customers with continuous service.’
With calls now being recorded and analyzed, the group can better identify its customers’ requests. According to Autotorino, about 70% of requests concern after-sales service, mainly booking workshop appointments for routine maintenance and vehicle drop-off. Callers also inquire about the status of their car in the workshop, which has led the group to work on a way to communicate this information to customers.
Eleven v4 and initial metrics
Autotorino tested Eleven v4 and notes new improvements. With the same voice, settings, and system prompt, the agent appears more natural and less like an automated system. The group highlights three developments: a more natural, less artificial voice; a more consistent and engaging tone that adapts to the stage of the conversation; and increased fluidity throughout the exchange. However, Autotorino wants the agent to remain direct and deterministic, so that callers get clear answers, while striving to make each conversation as natural as possible.
To track results, the group monitors three metrics: the number of leads created, the proportion of callers who hang up after hearing an agent but before a lead is created, and the sentiment analysis provided by ElevenAgents. According to the source, the hang-up rate has remained below the threshold set by Autotorino. The primary goal is to maintain a lead volume equivalent to the call volume, so that no prospect is lost. At this stage, the group reports an automated call volume more than double that of the previous system.
Replacing the voice server from November
In November, Autotorino plans to replace the interactive voice response (IVR) system at its first agency with ElevenLabs voice agents.
