A dispatch from 2030: The death of online classifieds?

The year is 2030. AI assistants improved rapidly through the mid-2020s and became the default way people interacted with the internet.

It started subtly. In 2026, AI assistants began responding to property queries with reasonable accuracy. Early results were rough, text-heavy, and incomplete. But they improved quickly. By 2027, the major AI platforms had negotiated data-sharing agreements directly with real estate agents, bypassing the portals to build their own listing databases. Agents, who had long resented rising classifieds fees, gave the new entrants their full support.

As traffic to the major portals began declining, management blamed cyclical weakness. But as agent memberships fell, it became increasingly clear that the doom loop had begun: fewer listings made the portals less comprehensive, driving more consumers to AI assistants, making agents even less willing to pay. The flywheel that had powered these businesses for two decades was spinning in reverse. Platforms increased marketing and technology costs as they fought to remain relevant. The 60-70% operating margins that had made these companies investor darlings were a distant memory.

But it is not 2030. Inspired by Citrini Research’s post on AI-driven white-collar job displacement, we have performed a similar exercise for dominant property classifieds such as Australia’s REA Group, Rightmove in the UK, Scout24 in Germany, and Hemnet in Sweden. These companies have seen their share prices fall over 30% since mid-2025 on fears of AI disruption. The European classifieds companies now trade on 12-17x next year’s consensus EPS, which implies a dire long-term scenario given that they continue to grow.

What would need to be true

The scenario above is internally coherent, but we believe it is highly unlikely. Each step depends on several low-probability assumptions. If any of them breaks, the narrative unravels.

First, AI platforms would need to build a complete, real-time property database. This would require direct relationships with tens of thousands of agents —over 16,000 UK branches for Rightmove and 26,000 in Germany for Scout24 — most of which are small businesses with portal tools embedded in their workflows. This fragmentation is one of the reasons online classifieds enjoy network effects in the first place. In practice, AI companies are doing the opposite: partnering with the portals to improve their own services.

Second, consumers would need to change entrenched behaviour when making the most financially consequential decisions of their lives. A large share of portal usage is browsing: buyers often don’t know what they want until they see it, discovering preferences and making difficult trade-offs over weeks or months. Engagement behaviour is telling. In Sweden, Hemnet attracts visits from most of the population monthly, with users returning more than three times per week on average. In Australia, a record 13 million people visited realestate.com.au in November 2025 — over 60% of the adult population — spending 36 minutes per visitor. Across the globe, over 80% of traffic goes directly to the major property portals consumers’ know and trust, while referral traffic from AI platforms remains below 0.5%. Scout24 noted that this share declined in 2025 vs 2024.

Third, agents and vendors would need to defect without consequence. An agent who cancels their subscription doesn’t just save a fee — their listings disappear from the platform, increasing the risk of a lower sale price. Vendors are unlikely to find this attractive when the platform fee remains small relative to the transaction size. Agents with access to the platforms will gain share from those without. To overcome this prisoner’s dilemma, in 2013 a consortium of UK estate agents attempted a coordinated defection from Rightmove by launching OnTheMarket. More than a decade later, Rightmove’s market share is largely unchanged. This scenario is even less likely in Sweden and Australia, where it is the seller, not the agent, who pays for the listing.

We’ve seen this movie before

The idea that a powerful technology company will disintermediate specialist classifieds is not new. Google launched property listings in Google Maps in 2009 and experimented across the UK and Australia, only to shut the feature down by 2011. In the US, investors estimate CoStar has spent >US$3 billion building Homes.com to challenge Zillow, generating ~US$80 million in revenue against Zillow's >US$2 billion. Traffic, technology, and capital alone are not sufficient. What matters is a complete, trusted listing database and a monetisable relationship with the supply side — assets reinforced by network effects and built over decades.

The more likely future

It is 2030. AI has changed property search, but not in the way many feared. AI assistants help users refine preferences, broaden discovery, and personalise recommendations. But these improvements have been underpinned by proprietary data accumulated over decades by the portals themselves. The major AI companies partnered with the leading portals rather than replicate the expensive work of maintaining a real-time listing database. Agents continued to pay for access to demand. The model proved resilient.

For long-term investors, the question was never whether AI would change these businesses. It did. The question was whether that change would undermine the durability of their profits. We could be wrong, but for now, we believe the market’s imagination is running ahead of reality — and that imagination is what is creating the opportunity.

To stay up to date with all the insights from the team at Canopy Investors, click here.