AI for Property: From First Enquiry to Closing, Faster

Saturday afternoon. An enquiry arrives from a property portal on your agent's personal WhatsApp, asking whether the corner unit is still available and what the monthly instalment would be on a KPR (a home mortgage). The agent is on site with another buyer. The reply goes out on Monday morning, by which time the prospect is already talking to another project. Nobody was careless. There was simply no system holding that question until the agent had a free moment.

AI for property businesses pays off most when it is installed in the order a deal actually moves, rather than bought as a bundle of features. That order is: log every lead in a system the company owns, reply and qualify quickly, keep follow-up alive through to the site visit, then chase the KPR paperwork once the booking fee is paid. At each step AI handles the logging and the reminders, while the agent still does the selling and makes the decisions.

The sequence matters more than the tool

This year's market leaves little room for wasted leads. Primary-market home sales in the second quarter of 2026 were still down 2.36% on a year earlier, though that is far gentler than the 25.67% fall in the first quarter (Source: Bank Indonesia, 2026). The BI-Rate was held at 5.75% as of September 2026 (Source: Bank Indonesia, 2026). With fewer serious buyers around, every unanswered question costs more.

Agents abroad offer an honest warning. In a technology survey of property agents in the United States, 48% of respondents already used AI at least once a week, yet 35% said it had made no difference to their business (Source: Real Estate News reporting the National Association of REALTORS, 2026). Those figures describe American agents and may not hold here. The lesson travels anyway: using AI to write listings does not necessarily move a single deal. Results show up when AI sits where deals usually stall.

AI for property, from first enquiry to akad

Follow that Saturday enquiry from the start. Step one is capturing it. Portal leads generally land on the agent's phone: Rumah123's agent app, for example, gathers leads from WhatsApp, phone calls and email on one screen (Source: Rumah123, 2026). The volume is large, since the portal said it was receiving more than 500,000 property enquiries a quarter on average in early 2025 (Source: Investortrust quoting Rumah123, 2025). The catch is that when an agent changes firms, the chat history leaves with them. So the first automation is a plain one: every enquiry from portals, ads, Instagram and web forms is logged automatically in one place the company owns.

Step two is the first reply. An older cross-industry study in the United States sent test enquiries to 2,241 companies; only 37% replied within an hour and 23% never replied at all, even though firms that made contact within the hour were nearly seven times as likely to qualify the lead as those that waited one hour longer (Source: SmartCompany reporting the study by Oldroyd, McElheran and Elkington in Harvard Business Review, 2011). The study is dated and it measured lead qualification, which stops short of a sale, so treat it as a compass. At this stage an AI assistant answers first and politely, asks the basics, then hands the conversation to the agent once the person is ready for a serious talk.

Step three is qualification. Budget, cash or KPR, preferred location and move-in timing are enough to score intent and route the lead to the right agent. In the sales system we built for a property agency, this part covers spam screening, an intent score for each lead, automatic assignment to agents, and a board that shows follow-ups past their due date. Built by Agentiq for an anonymised Indonesian property-sales agency; all screens shown use fictional sample data.

Working this way is still uncommon, even in more mature markets: in the United States only 21% of agents used a CRM with AI-powered insights and 7% used chatbots to capture leads (Source: HousingWire reporting the National Association of REALTORS, 2025).

Step four is keeping follow-up alive. Rumah123 puts the average agent's time to close at one to two months (Source: Investortrust quoting Rumah123, 2025). Over that stretch, human memory easily loses out to newer leads. AI prepares reminders and draft messages based on the last conversation, proposes site-visit slots that fit the agent's calendar, and flags prospects who have gone quiet. The agent checks, sends and shows up on site.

After the booking fee: the bottleneck few people talk about

In Indonesia, closings tend to stall after the booking fee has been paid. KPR financed 70.05% of primary-market home purchases in the second quarter of 2026 (Source: Bank Indonesia, 2026). One bank's guide estimates the process at roughly 8 to 13 weeks from application to disbursement, with the first one to two weeks spent preparing documents (Source: OCBC Indonesia, 2025). That opening stage is the only one a sales team can control.

The file is not a thin one. One bank's sample checklist includes the identity cards (KTP) of the applicant and spouse, the family card, the marriage certificate, a tax number (NPWP) and the last three months of bank statements, plus a payslip and an employment letter for salaried staff (Source: BCA, 2026). AI can hold a checklist per buyer, send reminders about missing documents, and tell the KPR admin when the file is complete. Checking the contents and dealing with the bank stay with people.

This year brings one more reason to move quickly. The 100% government-borne VAT incentive for new, ready-to-occupy landed houses and apartments, which applies to the first Rp 2 billion of the price on units costing up to Rp 5 billion, requires a handover document dated no later than 31 December 2026 (Source: DDTCNews, 2026). A file that drags through November can cost the buyer that saving.

The last step is short. Because every stage is logged, the weekly report assembles itself and you can see where most leads drop out.

What this means for your business

Start with data. A prospect's payslips, bank statements and KTP forwarded through an agent's personal WhatsApp are personal data. Indonesia's Personal Data Protection Law classes personal financial data as specific personal data and requires written notification within 3 x 24 hours if its protection fails (Source: Law No. 27 of 2022 on Personal Data Protection, 2022). Moving those files into a company system with clear access rights reduces that risk and speeds up the admin work as well.

On cost, you don't have to build everything at once. As a benchmark, Agentiq's starting prices are Rp 2 million for an AI agent and Rp 15 million for a custom app (Source: Agentiq, 2026). The final figure depends on how many channels, projects and stages you want to tidy up. Begin with the step that leaks most often in your team, measure it for a month, then move to the next.

Where do you start this week?

Four things can be done before you buy any tool.

• Count last week's leads that arrived on agents' personal phones: how many made it into the office system, and how long each waited for a first reply.
• Write down the qualifying questions your best agent always asks. That is the raw material for the AI.
• Map which KPR documents most often arrive late, and who has been chasing them.
• Decide where a human must take over, such as price negotiation and discount promises.

If you would like to see what this flow could look like for your project or agency, tell us about your process through our contact page. The Agentiq team will reply with concrete recommendations, not a template.