Ten Questions With Franck Le Tendre, Co-Founder of Omny

AI-powered real estate search is a very big topic in the industry at the moment. Although there is a lot of discussion around it, it's still pretty rare on real estate portals around the world.

The fact that some of the biggest portal names around the world are taking a while to go through their experimental phases and put truly granular search in front of their users has been the catalyst for a new wave of entrepreneurs building AI-native challenger platforms.

One of those building a platform to change how people search for property is former Groupe SeLoger executive Franck Le Tendre (pictured below). We caught up with Franck to find out about Omny, a platform which has raised €750k in pre-seed funding to change how people search for a home in France...

What is the problem that Omny solves?

Buying a home has become paradoxically harder in the digital age. There are more listings than ever, yet buyers feel less certain. They browse five or six portals, set alerts everywhere, and still worry they’re missing the right property. On the other side, agents receive a high volume of inquiries, but many are unqualified or poorly matched.

We are optimising for relevance, not visibility. Omny turns buyer intent into qualified demand for agents. Instead of forcing users to filter through forms and checkboxes, we let them describe their lifestyle, constraints, and preferences in natural language. Our AI understands context, not just criteria and connects that intent to the most relevant inventory across the market.

What made you want to build an AI real estate search platform?

After years of leading operations at Groupe SeLoger (part of AVIV Group), I had a front-row view of how digital real estate marketplaces function both economically and structurally.

Portals were built in the early 2000s around a simple paradigm: aggregate listings, generate traffic, sell visibility. That model was incredibly successful. It digitised classified ads and created massive distribution leverage for agents. But over time, something changed. Inventory became abundant. Attention became fragmented, and users became overwhelmed.

Buyers don’t think in filters; they think in lifestyles. For the first time, technology could interpret nuance, trade-offs, and context expressed naturally. That felt like an inflexion point.

During my years in the industry, I also saw how much inefficiency exists in demand distribution. Agents receive high volumes of leads, but too many are weak, misaligned, or exploratory. They spend time filtering rather than advising. If AI can understand buyer intent better upstream, then demand can be qualified before it reaches the agent. That changes unit economics, productivity, and satisfaction on both sides of the marketplace.

I didn’t want to add AI to an old architecture. I wanted to rebuild the architecture around intent.

What is the business model, and who are your customers?

We operate a multi-sided marketplace. On one side are the buyers and renters, who access Omny for free. On the other side are the real estate agents and networks.

Our model is pay-per-lead, powered by a token system. Agents only pay to unlock qualified buyer contacts, and pricing varies depending on the scoring and strength of intent. We do not charge for simple visibility. We monetise qualified demand.

Over time, we also integrate services around the housing journey, like financing, renovation, and home equipment, through partnerships.

How does Omny get its inventory?

From day one, we chose not to depend on scraping portals or reselling aggregated listings. Instead, we built direct supply connectivity via connections with agency CRMs and agency websites.

We use something called SmartSync, which connects directly to the agent’s own digital infrastructure, not third-party portals. That means we synchronise from the original source, which allows us to access 100% of participating agencies’ listings and preserve data ownership and attribution.

Strategically, this approach allows Omny to federate supply across the market without forcing agents to duplicate publication workflows. Over time, this creates something very powerful: a federated, AI-indexed layer of professional supply connected directly at the source.

French real estate listings don’t seem to have that many fields of information. How do you make sure that if a user asks for features like ‘south-facing garden’ or ‘off-street parking’ in their request, you don’t deliver false positives?

You’re absolutely right, structured data in real estate is often incomplete.

Most listings were not designed for semantic search; they were designed for human reading. As a result, key information is often missing from structured fields, buried inside long descriptions, implicit rather than explicit, or visible only in photos. If we relied solely on traditional database filters, the experience would be unreliable. So we built a multi-layered data stack.

First, we analyse structured fields, unstructured descriptions, and images through semantic and visual AI.

Second, every listing is enriched through connections to 50+ external datasets, including for instance, seismic and flood risk, clay shrink–swell exposure, demographics, transport, education, healthcare, shops and services, urban zoning information, environmental indicators, fiber coverage, etc…

This allows us to match intent against contextual intelligence, not just keywords. We're also continuously learning from user behavior like saved properties, refinements, and engagement depth.

Traditional portals match fields. Omny matches intent + context + environment. That’s a fundamentally different architecture and it’s why conversational search in real estate only works if you build the data stack underneath it.

You were previously VP of Operations at Groupe SeLoger. How did that experience prepare you to launch Omny?

My time at Groupe SeLoger (part of AVIV Group) gave me a front-row seat to how large-scale real estate marketplaces truly operate beyond the surface metrics. I worked closely with product, sales, data, and network partners. I saw how lead flows are generated, how agents evaluate ROI, how pricing models influence supply behaviour, and how marketplace dynamics evolve when competition intensifies.

Agents were paying more for visibility while struggling with declining lead quality. Buyers were browsing more listings but feeling less confident in their search. Traffic was growing, yet satisfaction wasn’t necessarily improving.

That experience helped me understand something fundamental: the traditional portal model is structurally optimised for volume — page views, listings, impressions. But the future belongs to platforms optimized for match quality and intent understanding.

It also gave me credibility with professionals. I understand their economics, their constraints and their scepticism toward “new shiny tools.”

Omny was not built from a theoretical perspective. It was built from operational reality. In many ways, I understood that the industry doesn’t need another portal. It needs a structural shift in how demand is qualified and distributed.

How does Omny plan to build traffic and scale the business?

We approach growth structurally. First is SEO. We build high-intent content around real estate, financing, regulation, pricing, and AI-driven discovery. This compounds over time.

Second, paid acceleration (SEA and social). We test high-intent acquisition while maintaining performance economics aligned with our pay-per-lead model.

Third, product-led growth. Real estate decisions are social, and Omny enables collaboration and sharing, creating organic distribution loops.

Fourth, MCP integration. As conversational AI platforms become primary interfaces, Omny is developing a connectivity layer allowing professional listings to surface directly within AI-native search environments.

Traffic will increasingly come from conversations, not just browsers. Scale does not come from buying impressions. It comes from becoming the intent infrastructure.

Which products or companies have you taken inspiration from when building Omny?

Inspiration came from very different places, not necessarily from real estate.

From a UX perspective, platforms like Lovable and Suno inspired us with their immediacy. You type, and something meaningful happens. The interface feels invisible. Real estate search should feel fluid, not bureaucratic.

From an adoption perspective, I admire Dropbox, Notion, and Tesla. They didn’t just improve products, they reframed expectations. Dropbox made storage invisible; Notion unified fragmented workflows; Tesla redefined a category through software and design.

Omny follows similar principles:

• The first interaction must deliver value instantly.
• The product must feel modern enough to signal “this is different.”
• Sharing must be natural — because real estate decisions are social.
• Adoption should not require education; it should feel intuitive.

We benchmark against the best consumer apps — not against traditional real estate portals.

Is there proprietary data behind Omny’s search or does it rely on agents being accurate and detailed with their descriptions?

Yes, and that’s where our moat emerges.

We start with agent listings, but we enrich them through 50+ external territorial datasets. That creates a contextual intelligence layer around every property. This means that every listing becomes more than a description; it becomes a contextualised real estate asset. When a buyer says “I want a family-friendly neighbourhood”, or  “I want a dynamic area with shops” We’re not guessing. We’re matching against structured territorial intelligence. This enrichment layer already makes Omny fundamentally different from simple listing aggregators.

More importantly, we build a continuously learning intent layer. We analyse conversations, refinements, saved and rejected listings, engagement depth, and lead conversion signals. Over time, this creates structured intent intelligence across cities and territories. That intent layer compounds. And unlike listings, which are public and cyclical, intent data is cumulative.

A competitor could copy the interface. They could even access similar listings. But replicating years of structured intent data combined with enriched territorial intelligence is exponentially harder.

What is one thing you think the real estate marketplace industry should be talking about more?

Lead quality, transparency and buyer experience quality.

The industry talks endlessly about traffic volumes and listing counts. But what truly matters is not how many people click, it’s how many meaningful matches happen. For agents, that means conversion probability. If a majority of inquiries are weak or exploratory, volume becomes noise.

But for seekers, the problem is symmetrical. Buyers spend months browsing, comparing, worrying they’re missing something better. They receive alerts that don’t really match. They visit properties that look right on paper but feel wrong in reality.

Low-quality demand hurts agents, and low-quality matching exhausts buyers. We should be discussing intent scoring, conversion tracking, match quality metrics and performance accountability.

And from the buyer’s perspective, reduced search fatigue, fewer irrelevant visits, and higher confidence that “I’m not missing out”.

Real estate marketplaces need to move from selling exposure to selling probability for both sides. Because ultimately, the real metric isn’t traffic, it’s confidence. And that is what makes people move forward