An AI agent for social housing sits upstream of the contact centre. It captures the tenant's request, qualifies it, then either resolves it itself or hands it over to the right advisor with its full context. A social housing provider's chatbot answers a question; the agent decides what should happen to it.
DialOnce calls this the AI agent for contact journeys: a single agent, present on the phone, in the tenant portal, on the website and in the mailbox, with the same understanding of the request whatever the channel.
The agent picks up the request wherever the tenant tries to reach their provider (online tenant portal, phone, QR code...). It then identifies the real reason for contact, beyond the words used. When it has the answer and an advisor would add nothing more, it resolves the request by giving the expected answer. Otherwise it hands over, and on the right channel: an administrative matter can wait, a water leak cannot.
The agent is trained on more than 550 contact intents specific to social housing. They follow the tenant lifecycle, from moving in to ending the lease, through technical requests, account questions and administrative procedures. This scope covers almost all of a housing provider's contact volume.
Every channel leads to the same desk, with no prior sorting. Advisors receive raw contacts, tenants who get no answer try again elsewhere, and the volume swells.
Over the years, organisations have multiplied their contact points: a phone number, an email address, a form, sometimes a chat. Tenants do not really choose their channel. They use whichever is at hand and start again elsewhere if they get nowhere.
The advisor then has to piece the situation back together before handling it. Meanwhile, the tenant sends another email, then calls back. Repeat contacts inflate the volume, the volume puts the team under pressure, and reachability drops.
This was Adrien Lesage's starting point at the Pitch Zone. The contact centre only sorts a request once it has entered the system, and by then it is too late to act on the volume. Everything happens before. Tenant expectations point the same way, as we explained in our article on how AI is transforming tenant relations in social housing.
A tenant calls because they do not understand the service charge adjustment shown in their tenant portal. The AI voice agent answers and explains that the adjustment reconciles each year the advance payments with the actual costs. No advisor steps in, because the request did not need one.
The exchange feels like a real conversation. The tenant follows up with a question about what those actual costs cover. The agent answers, then offers to guide them through the tenant portal to find the details. The tenant has what they wanted and hangs up.
The same call about a water leak goes differently. The agent recognises the emergency and transfers the tenant to an advisor without keeping them waiting (through a live chat with a person or by displaying a priority phone number), along with the context already gathered. The advisor does not need to ask the questions again.
One point could not be heard in the demonstration, which was run in French: the agent handles every language. For teams dealing with strong linguistic diversity across their housing stock, the benefit is immediate.
No. The agent takes care of repetitive requests, the ones it can resolve on its own. Advisors keep their time for the situations that deserve it: a tenant in difficulty, a sensitive case, a conversation that calls for listening. In short, advisors do what they do best: support people with human sensitivity.
It was the question we were asked most often on the stand, by business leaders and IT directors alike. It is understandable in a sector where the social mission depends on human contact. The agent does not take that contact away from tenants; it keeps it for the moments that matter.
Brest Métropole Habitat approached the subject from the other end, by equipping its advisors directly. The provider manages nearly 20,000 homes, and its contact centre of 10 advisors handles 100,000 calls and 10,000 emails a year. With its Augmented Customer Advisor project, advisors can access scripts and the knowledge base during calls, and incoming emails are sorted and prioritised. The same logic runs through our 5 concrete use cases of AI in social housing.
"The goal of the Augmented Customer Advisor project is to give our team access to our internal scripts and news in one click or with a simple question, without interrupting the conversation with our tenants. A real gain in peace of mind and efficiency." Lénaïck Colin, Quality and Internal Control Manager, Brest Métropole Habitat
Across all DialOnce customers, an agent placed upstream of the contact centre avoids up to 25% of contacts and improves reachability by up to 30 or even 35 points. For tenants, waiting disappears on simple requests, including at night and at weekends.
Advisors benefit too. They receive fewer requests, and those requests arrive qualified and in context. The time saved goes to situations that call for judgement and empathy, and social housing has no shortage of them.
At 1001 Vies Habitat, an inbound call can switch to a digital journey, where the request is qualified more precisely and then resolved autonomously.
"Thanks to DialOnce's conversational agent, we have not only improved the accessibility and responsiveness of our service, but also optimized the management of incoming requests, allowing our tenants to get instant answers to their queries, 24/7." Maud Flory-Boudet, Head of Multichannel Customer Relations, 1001 Vies Habitat
Read the 1001 Vies Habitat customer story and see how the provider deployed its AI agent for its tenants.
The simplest way is to start from your own contact reasons rather than a generic demonstration. A 45 to 60-minute workshop with your teams is enough to estimate what can be resolved autonomously.
The workshop runs in four stages:
You leave with a first matrix that cross-references contact reasons, volumes and resolution potential.
Yes. The agent handles every language, which helps teams facing strong linguistic diversity among their tenants.
Yes. Everything that can be resolved autonomously is resolved at night and at weekends too, without involving the teams.
No. The DialOnce solution is technology-agnostic and integrates with the main CCaaS and CRM platforms on the market, with official partners such as Kiamo. DialOnce feeds your contact centre with qualified, prioritised and contextualised requests, without replacing it.
In France. DialOnce relies on a proprietary engine designed and hosted in France. It is a sovereign solution, developed by a French software company that has supported more than 150 large organisations for ten years.