Key points
- Immobiliare.it, Italy's largest real estate platform, built on ElevenAgents, starting with a voice agent on property listings and expanding to a network of agents.
- Over 70% of sellers opted in to the first agent, which answers buyer questions, collects lead details and summarizes each conversation for real estate agents.
- Later agents handle outbound calls to improve listing quality and conversational search; subagents split work behind a single conversation.
- The company pursued AIUC-1, an emerging certification standard for conversational agent security, safety and reliability.
Why it matters for marketers: A listings marketplace shows how a single lead-qualification voice agent can grow into coordinated specialist agents, with evaluation and guardrails built in from the start.
Immobiliare.it, which lists more than 1.2 million properties in Italy and serves over 31,000 agents and developers, started with one voice agent attached to each listing. It answered buyer questions at any hour in the right language, gathered contact details and wrote a summary so real estate agents could judge how promising each lead was. According to ElevenLabs, more than 70% of sellers opted in.
The company also reports that buyers seemed more at ease giving their information in a spoken exchange than in a typed one, which it says improved lead quality. From that base, it added an outbound agent aimed at improving listing quality and a conversational search tool for browsing properties.
As the number of agents grew, the team moved to subagents: small, specialized agents that each take one narrow job while the user experiences a single conversation. Paolo Sabatinelli, the company's chief product officer, said much of the innovation lies in orchestration and keeping behavior predictable. The team also pursued AIUC-1, and ElevenLabs says the platform's built-in guardrails covered most of that standard.
For marketers and content teams, the case is a pattern more than a product pitch: start with one narrow, measurable conversation, learn from transcripts, then split the work into specialized pieces. It also shows lead quality and opt-in rates can serve as the metrics that justify expansion.
Worth watching is how widely AIUC-1 is adopted, and whether other marketplaces copy the subagent approach. Note that the results come from a vendor case study, so figures such as the opt-in rate have not been independently verified.
Sources
This brief is our own summary. Read the full reporting at the source.