Case study
Living sector operator in Madrid · AI agent "Eva"
Back to case studies
Living and residential rental Madrid, Spain

When growing means more enquiries than you can answer

Here is how Nuva Living solved it with Eva, an AI agent that answers the prospect, books the viewing, and follows up until the team steps in.
Nuva Living – Madrid, Spain
A living sector operator in Madrid. The operating platform of Medan Asset Management, backed on some of its assets by institutional investors such as Partners Group and Invesco. 1,500 units in operation and around 1,500 more in the pipeline, across Madrid and its metropolitan area, under two brands: Nuva Home for long-term rentals, and Nuva Flex for medium-term stays of 1 to 12 months with services and shared amenities. nuvaliving.com

In just a few years, Nuva Living has become one of the most active living sector operators in Madrid. Buildings in Valdemoro, Atocha, Tetuán, and Torrejón de Ardoz, two stay formats, and a digital operating model designed to make renting feel more like a service than a contract.

That growth has a direct consequence: every new building multiplies the enquiries. How big is the studio, what does the rent include, are pets allowed, when can I visit, is anything free for next month. Legitimate questions, repeated constantly.

An IVR was there to filter them. A menu of options that could not work out which development the caller was asking about, nor resolve a routine question, so the conversation ended up with the team anyway. Between 30% and 35% of their working day went into it.

The decision was not to hire more people to answer the same things more times. It was to look for a structural solution that could scale at the same pace as the portfolio.

That is how Eva was born: an artificial intelligence agent that handles prospect enquiries, books and rebooks viewings, follows up with reminders, and recognises anyone who has called before. And when a conversation needs a person, it alerts the team with the full context of what was said.

This is the story.

The company

The starting point

Nuva's operation worked. But as lead volume grew, the same bottlenecks appeared that any commercial team will recognise:

An IVR that filtered nothing. The menu could not identify which development the caller meant, nor answer the usual questions, so almost every call reached the team. Between 30% and 35% of their daily time.
Lead volume that kept climbing. Every opening added enquiries, and the system they had did not scale at the pace of the business.
Out of hours, nobody on the other end. The IVR picked up, but the enquiry stayed unresolved until someone on the team called back. In many cases, by then the lead was already gone.
Viewings arranged one call at a time. Offering a slot, confirming it, and reminding for every viewing consumed time that was not spent closing.

This is not exclusive to the living sector. Any company whose team handles a high volume of repetitive enquiries before a sale even exists has the same challenge: estate agencies losing prospects out of hours, clinics where reception books appointments instead of caring for patients, hotels answering the same ten questions every day, or dealerships coordinating test drives. The sector differs. The problem is the same.

The solution

The decision

With Telvia they built Eva: an AI agent that handles prospect enquiries through whichever channel they use, with the real information for each building and each home type. It is not another menu of options, nor a chatbot with canned answers: it works out which development they are asking about even when they do not know its name, answers the specific question they came with, and carries the conversation. It informs in detail, offers a viewing slot, books it, moves it if the prospect needs a different time, sends the reminder, and alerts the team internally when a person is needed.

Voice
WhatsApp
Email

What Eva is able to do

Resolves the enquiry

Home types, sizes, what the rent includes, services and shared areas, stay conditions, availability. At whatever level of detail the prospect asks for, with no waiting for a callback.

Books, confirms, and rebooks

Offers available slots, arranges the viewing at the building that fits, and logs it. If the prospect needs to move it, they rebook it themselves without going through the team.

Handles follow-up

Reminders to the prospect ahead of the viewing, and follow-up on conversations left open. Fewer no-shows, fewer leads going cold on their own.

Alerts the internal team

An internal alerting system: if the conversation goes off-script, if the lead is ready for the next step, or if a commercial query needs intervention, the team knows immediately, with the full context of what was said.

Answers 24/7

A Saturday night, a Sunday morning, or a Tuesday at seven. The enquiry is handled the moment the prospect is interested, not two days later.

Recognises returning callers

If the prospect has spoken to Nuva before, Eva knows who they are, what they asked last time, and whether they already have a viewing booked. It never starts from scratch or asks for their details again.

From email to WhatsApp

It filters incoming emails from prospective renters and carries the conversation over to WhatsApp, where the answer arrives immediately instead of sitting in an inbox.

Handles the commercial query

Questions that come up during the process, changes to a booked viewing, or a follow-up left hanging. Commercial matters only: it does not cover maintenance or the running of the home.

Logs whatever it cannot close

Eva does not hand the call over live. When something is left unresolved, it logs the conversation with full context — who they are, what they asked, and where they got to — so the team can follow up without the prospect repeating anything.

Before vs. now

Unresolved
Until someone called back
Instant
Enquiry resolved and viewing booked

The turning point

A prospect calls on a Saturday night. Before Eva, the IVR picked up and walked them through a menu that did not answer their question. On Monday someone returned the call and, with luck, the person was still interested. In many cases they had already signed somewhere else.

That is the lead that never shows up in any report: it is not lost on price or product, it is lost to a 48-hour gap between the moment someone decides they are interested and the moment someone can reply. It is also the cheapest one to win back, because it requires changing nothing about the business. Just being there.

The rollout

How the process went

Nuva Living brought the business knowledge. Telvia designed the agent's architecture, connected it to the operation, trained natural language understanding, defined the alerting and logging rules, and took it to production.

1
Map the real conversations

Before writing a single line of agent, understand what a prospect actually asks, in what order, and in how much detail. The conversations the team was already having were the best raw material.

2
Build the knowledge base

Buildings, home types, services, stay conditions, and availability, structured so Eva could query them in real time. If the agent was going to talk to real prospects, it could not get a condition or a property detail wrong.

3
Wire up scheduling, reminders, and alerts

Eva informing is good. Eva also closing the viewing in the calendar, sending the reminder, and raising the internal alert is what turns the agent into part of the operation instead of one more channel someone has to check.

4
Production, call reviews, and continuous improvement

Eva went live and Telvia stayed on top of it: listening to real calls, monitoring how the agent handled each conversation, and tuning its understanding case by case. Improvements are proposed and reviewed together with the Nuva team, never applied blind. It is the part Nuva highlights about working together.

What Nuva Living brought
An owner who knew the lead journey in detail
The product knowledge: buildings, home types, conditions
The criteria for when a person should step in
Access to the viewing calendar and the real operation
Continuous feedback on Eva's conversations
What Telvia built
An AI agent for lead handling and qualification
A knowledge base queryable in real time
Contact recognition and history with Nuva
Automatic viewing scheduling, confirmation, and rebooking
Email filtering for prospects and handover to WhatsApp
Commercial query handling and internal team alerts
Call reviews, monitoring, and joint continuous improvement
The outcome

What changed

Most conversations never reach a person. And the ones that do, arrive at the best possible moment.

80-90%
Conversations resolved with no human involved
30-35%
Of the team's daily time, recovered
24/7
No leads lost out of hours

The team no longer spends a third of the day answering the same things. It now steps in when the lead is further along, at the viewing, which is where a person adds something an agent cannot.

And there is an effect that was not in the original plan: Eva gives more thorough information, faster than a person juggling ten conversations at once can. That, according to Nuva, improves lead conversion. Not because it sells better, but because it answers sooner and answers in full.

Working together

Why Telvia

What surprised the Nuva team most was not the technology. It was the level of attention: reviews of real calls, continuous monitoring of how the agent responded, and improvements put on the table before anyone asked for them, worked through together with the Nuva team.

Who ran the project at Nuva
Jaime Morán
Jaime Morán
Founder & CEO · Medan Asset Management
Valentino Chiavacci
Valentino Chiavacci
Investment Analyst · Medan Asset Management
Valeria García
Valeria García
Marketing Manager · Medan Asset Management
Jorge Marcos
Jorge Marcos
Product Operations Manager · Medan Asset Management

"The Telvia team is always on top of the project, proposing improvements proactively. It is not just a tool, it feels like having a partner making sure everything works."

Valentino Chiavacci
Investment Analyst · Medan Asset Management
Guarantees

Integration and security

Connected to the operation. The agent works with the real business data: availability, home types, conditions, and the viewing calendar. The integration works with any CRM, PMS, or system with an API.
Encrypted communications. All traffic between the agent, the channels, and the client's systems travels encrypted. Data at rest is stored encrypted too.
Full isolation. Nuva Living's data is never mixed with any other client's. It is not used to train external AI models.
Regulatory compliance. Aligned with GDPR and Spain's LOPDGDD, with a processing notice inside the conversation and full traceability of every action the agent takes.
Does your company have a similar challenge?
Your team spends a large share of the day answering the same enquiries
You lose leads because nobody answers out of hours
Coordinating viewings or appointments eats into commercial time
You need to scale without adding headcount proportionally
Tell us about your case in 30 minutes

"It is a useful tool and it is worth implementing: it frees up the team's time and improves customer service without losing quality."

Valentino Chiavacci
Investment Analyst · Medan Asset Management