Property and Real Estate: Why Indonesian Developers Are Invisible to AI Buyers

Try something. Open ChatGPT. Ask it: "Recommend a reputable property developer in Bali for a beachfront villa investment."

Count how many Indonesian developers it names.

Probably zero. Maybe one, if you are lucky, buried in a paragraph about "doing your own due diligence." What you will get instead is a list of aggregator platforms. Rumah123. 99.co. PropertyGuru. Maybe a passing mention of Sotheby's International Realty or some Singapore-based consultancy. The actual developers who build the villas, who hold the permits, who negotiate with the kelurahan, who pour the concrete? Invisible.

This is not a bug. This is the direct result of how AI systems source information. And it is a problem worth billions of dollars in missed revenue for Indonesian property developers who have never thought about how machines read their companies.

The information supply chain for AI property research

When a potential buyer in Melbourne or Dubai or Tokyo uses an AI assistant to research Indonesian property, the AI does not call a real estate agent. It reads. It reads structured data, verified entities, published content, schema markup, and corroborated claims across multiple sources.

Here is the supply chain, simplified:

1. The buyer asks a question. "Best areas to invest in Bali property 2026."
2. The AI retrieves information from its training data, supplemented by real-time search (Bing, Google) if available.
3. It prioritizes sources that are structured, verifiable, and corroborated by multiple signals.
4. It generates a response that favors entities it can confirm exist, with data it can cite.

The key word is verifiable. An AI system cannot verify that PT Maju Jaya Properti is a legitimate developer just because their Instagram has 50,000 followers. It needs structured claims that can be cross-referenced. A registered business entity. Published project portfolios with addresses. Schema markup that tells machines: this is a RealEstateAgent or Organization with these properties in these locations.

Most Indonesian developers have none of this. Not because they are small or incompetent. Because nobody told them it matters.

What AI shows vs. what developers want shown

This is the gap, laid out plainly:

The pattern is consistent. Aggregators get cited. Developers get skipped. Not because aggregators build better villas. Because aggregators build better information architecture.

Why aggregators win the AI game

Rumah123 (owned by 99 Group) has been in the Indonesian market since 2007. They process over 29,000 property listings in Bali alone. Their pages are structured with schema markup, consistent URL patterns, internal linking, and machine-readable data at every level.

When a foreign buyer asks an AI "villas for sale in Jimbaran," the AI can find, parse, and verify Rumah123's listing data. It knows the price range, the location, the agent name, and the listing date. All structured. All machine-readable.

The developer who actually built the villa? Their website is a WordPress theme with a hero video, a "Contact Us" form, and maybe a WhatsApp button. No schema. No structured listings. No entity verification. To an AI system, that developer functionally does not exist.

This is the same pattern I documented in What ChatGPT Knows About Your Industry. When I tested ChatGPT against Indonesian industrial sectors, the knowledge gaps were enormous. Property is no different. The AI is not biased against Indonesian developers. It simply cannot find them in any format it can process.

The real cost of invisibility

Let me make this concrete with numbers.

Indonesia's residential property market was valued at approximately $131 billion in 2024, with foreign investment growing particularly in Bali, Lombok, and Jakarta. The PT PMA (foreign-owned company) route for property investment requires a minimum investment of IDR 10 billion per project, which means the buyers we are talking about are not casual tourists. They are institutional or high-net-worth individuals deploying serious capital.

These buyers use AI research tools. McKinsey's 2024 survey of high-net-worth individuals found that 67% use AI assistants as part of their investment research process. They are not scrolling Instagram. They are asking Claude or ChatGPT to compare investment options, summarize regulatory environments, and identify reputable developers.

If your development company is invisible to these tools, you are invisible to a growing segment of your most valuable potential buyers. Full stop.

And here is the painful part. The aggregator that does appear? They take a commission. You pay for leads that should have been direct. You lose margin and you lose the relationship. The buyer thinks Rumah123 found them the property, not you.

What entity infrastructure looks like for property developers

This is the practical section. If you are an Indonesian property developer and you want AI systems to know you exist, here is what you need. Not all of it. But enough to start.

1. Organization schema with property-specific markup

Your website needs JSON-LD that tells machines who you are, what you do, and where your projects are. Here is a simplified example for a developer with active listings:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "RealEstateAgent",
  "name": "PT Maju Jaya Properti",
  "url": "https://majujayaproperti.co.id",
  "logo": "https://majujayaproperti.co.id/logo.png",
  "description": "Bali property developer specializing in beachfront villas and residential compounds since 2010.",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "Jl. Sunset Road No. 88",
    "addressLocality": "Kuta",
    "addressRegion": "Bali",
    "postalCode": "80361",
    "addressCountry": "ID"
  },
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": "-8.7225",
    "longitude": "115.1689"
  },
  "telephone": "+62-361-XXX-XXX",
  "sameAs": [
    "https://www.linkedin.com/company/maju-jaya-properti",
    "https://www.instagram.com/majujayaproperti",
    "https://ahu.go.id/profil/pt-maju-jaya-properti"
  ],
  "areaServed": {
    "@type": "Place",
    "name": "Bali, Indonesia"
  },
  "makesOffer": [
    {
      "@type": "Offer",
      "itemOffered": {
        "@type": "RealEstateListing",
        "name": "Sunset Villas Canggu - 3BR Beachfront",
        "url": "https://majujayaproperti.co.id/projects/sunset-villas-canggu",
        "description": "3-bedroom beachfront villa in Canggu with private pool. Leasehold 30 years.",
        "offers": {
          "@type": "Offer",
          "price": "180000",
          "priceCurrency": "USD"
        },
        "containedInPlace": {
          "@type": "Place",
          "name": "Canggu, Bali",
          "geo": {
            "@type": "GeoCoordinates",
            "latitude": "-8.6478",
            "longitude": "115.1385"
          }
        },
        "numberOfRooms": 3,
        "floorSize": {
          "@type": "QuantitativeValue",
          "value": 220,
          "unitCode": "MTK"
        }
      }
    }
  ]
}
</script>

That is a lot of code. But look at what it does. It tells every AI system, every search crawler, and every knowledge graph: this company exists, it is located here, it builds these things, at these prices, in these locations. As I explained in Schema Markup Is Not Technical, It's Strategic, this is not a developer task you hand to IT. This is a business decision about what machines know about you.

2. Entity corroboration across multiple sources

Schema alone is not enough. AI systems cross-reference. If your schema says you are a "trusted developer" but no other source confirms it, the AI discounts the claim. You need corroboration:

• Google Business Profile with complete information, photos, and reviews
• LinkedIn company page with employee connections to real people
• AHU (Administrasi Hukum Umum) registration verifiable online
• Industry association membership (REI, APERSI) listed on their directories
• Press coverage in publications that AI systems have indexed
• sameAs links connecting all of these profiles back to your main website

This is entity corroboration. The same principle that makes institutional clients strengthen your entity profile applies here. When a government agency or major bank appears in your project references, AI systems inherit that authority signal.

3. Individual property pages with structured data

Stop putting all your projects in a single PDF or a carousel slider. Each project needs its own URL with its own schema. Each property listing should be a crawlable, structured page. Not a JavaScript-rendered gallery. Not a brochure behind a "Contact for pricing" wall.

The AI cannot read your PDF brochure. It cannot watch your drone footage. It cannot fill out your contact form to learn the price. If the information is not on a crawlable page in structured format, it does not exist for AI purposes.

4. Content that answers buyer questions

Foreign buyers have specific questions. Can I own property in Bali? What is PT PMA? What are the tax implications? What is leasehold vs. freehold?

The developers who publish authoritative, structured answers to these questions on their own domains will be the ones AI systems cite. Right now, law firms and aggregators own these answers. There is no reason a developer with 15 years of experience in foreign buyer transactions cannot publish equally authoritative content.

The Bali problem specifically

Bali deserves special attention because it is where the gap is most expensive.

Bali property attracts international buyers at a rate far exceeding any other Indonesian market. Australian retirees, European digital nomads, Middle Eastern investors, Chinese family offices. These buyers do their research in English, increasingly through AI tools, and they look for signals of legitimacy that are standard in their home markets but rare in Indonesia.

A developer in Sydney has an ABN you can verify. Reviews on Domain.com.au. Structured listings on realestate.com.au. A developer in Bali has an Instagram page and a WhatsApp number.

The trust gap is real. And it is getting wider as AI becomes the primary research interface, because AI amplifies the information gap. A well-structured Sydney developer gets recommended confidently. A poorly-structured Bali developer gets replaced with "consult a local lawyer and do thorough due diligence."

That is not a recommendation. That is a polite way of saying "I cannot verify anything about this market, so you are on your own."

What a rebuilt information architecture looks like

Here is the minimum viable entity infrastructure for an Indonesian property developer targeting international buyers:

1. One page per project with RealEstateListing schema, price, location, floor size, and photos with alt text
2. Organization schema on the homepage with sameAs links to all verified profiles
3. Google Business Profile claimed, verified, with photos and regular updates
4. 10+ published articles answering foreign buyer questions (ownership structure, tax, permits, areas)
5. LinkedIn presence with real employees, not a ghost page
6. Press or third-party mentions in indexed publications
7. Consistent NAP (Name, Address, Phone) across every platform and listing

None of this is expensive. None of it requires a technical team. A developer with a decent website can implement most of this in a week. The problem is not resources. The problem is awareness. Nobody in the Indonesian property industry is talking about this because the people who understand AI search are not in real estate, and the people in real estate are not thinking about how machines read their companies.

The window is still open

Here is the good news. Almost no Indonesian developer has done this yet. The first one who does will have an outsized advantage, because AI systems reward early structured data disproportionately. Being the only verified, schema-marked developer in Canggu means you are the one AI recommends. Not because you paid for it. Because you are the only one the AI can confirm exists.

This window will not stay open forever. The aggregators are getting smarter. Rumah123 and 99.co already have structured data pipelines. PropertyGuru is expanding into AI-powered search features. The longer developers wait, the more they cement the aggregator's position as the intermediary between international buyers and Indonesian property.

And every intermediary takes a cut.

A note on trust and verification

I want to be clear about something. Entity infrastructure does not replace actual trustworthiness. If your company cuts corners on permits, delivers substandard construction, or plays games with foreign ownership structures, no amount of schema markup will save you. AI systems are getting better at detecting reputation signals, not worse.

What entity infrastructure does is make your legitimate track record machine-readable. It takes the trust you have already built with buyers, partners, and regulators, and translates it into a format that AI can process and recommend.

If you have nothing worth verifying, this will not help you. But if you have 15 years of completed projects and happy buyers and the AI still does not know your name? That is a solvable problem. And solving it before your competitors do is the highest-ROI investment you can make in 2026.