【Entering the Era of AI Negotiating with Real Estate Companies】FANGO Releases Japan's First AI Agent-to-AI Agent Real Estate Consultation Portal 'fango.city'

FANGO Inc. (Headquarters: Chiyoda-ku, Tokyo; Representative Director: Mei Goto) today released the beta version of 'fango.city', a real estate consultation portal where AI agents on the user side and AI agents on the real estate company side directly 'negotiate' to find the optimal property.
This is the first initiative in Japan for a portal where AI agents negotiate properties with each other (according to the company's own research).

This service is based on the company's patent-pending technology, combining a three-layer architecture consisting of Customer-side AI, Central Orchestration AI, and Agent-side AI, with blockchain-based action history management.

From the era of humans browsing real estate sites to an era where AI agents tirelessly search for the optimal property.
FANGO provides the infrastructure to support this transition.

The End of the Era of 'Humans Searching' Real Estate Information Sites

"By 2026, query volume on traditional search engines will decline by 25%, and search marketing share will be taken by virtual agents such as AI chatbots."
This forecast from 2024 by US research firm Gartner is now becoming a reality.

In the US, Google's search volume per person has fallen approximately 20% year-over-year.
Global search share also recorded its largest annual decline in the past decade.

Meanwhile, traffic from AI agents has surged [up 7,851% year-over-year].
According to the same report by HUMAN Security, automated traffic is increasing about eight times faster than human traffic.

Already, the Main Players in Web Access are 'AI and Bots'

In 2024, for the first time in 10 years since measurements began, the majority of all internet access was automated communications like bots, not humans, surpassing human access numbers.
This reversal was driven by the rapid spread of AI and Large Language Models (LLMs).

Furthermore, almost all AI returns answers without clicking links on external sites, failing to send users to those sites.
It must be said that the outlook for existing models that have relied on advertising revenue is becoming increasingly severe.

Moreover, by 2026, agent platforms like OpenClaw and Hermes Agent, which operate autonomously in personal local environments, have also emerged.
These AIs access users' personal information, understand their preferences with a resolution that even the users themselves are unaware of, and tirelessly search for properties.

This is no longer a story of the 'near future'.
The web has already become a place where more machines (AI and bots) traverse than humans.
The very act of a person opening a real estate information site and searching for properties with their own hands is now becoming a thing of the past.

Reaching the Optimal Property by 'Negotiating' AI-to-AI

It's not just end-users who use AI.
A world is already becoming a reality where AI agents that deeply understand the intentions of the principal stand on the side of the lender/borrower or seller/buyer.

The next stage, which FANGO is certain of, is a property information provision mechanism where 'a personal AI agent and a real estate company's AI agent directly negotiate with each other,' rather than interactions between human customers and human representatives of real estate brokerage firms.

The two AIs negotiate repeatedly, and as a result, a match with a property that meets the conditions is established.

A scene showing a personal AI agent and a real estate company's AI agent negotiating

Many off-market properties that are not publicly listed lie dormant in the hands (locally) of real estate companies.
This is because, until now, companies were forced to choose between 'letting them sleep without showing anyone, scattering information based on the subjective judgment of a representative, or uploading them to a central server and exposing them to an unspecified number of people.'

The AI era is different.
AI guides the owner and safely delivers information only to those who truly need it.

Properties circulating online are just the tip of the iceberg of the entire market.
FANGO's goal is not to replace existing real estate information sites, but to complement them, shine a light on properties that have been hidden beneath the surface, and expand a space where everyone can transact with peace of mind.

The Key is 'Trust' – Protecting Privacy and Eliminating Malicious Information

In a world where AIs negotiate properties with each other, new questions arise.
'How do we protect users' personal information and privacy?'
'How do we eliminate 'malicious information' such as fake property listings, impersonation, and tampering?'

According to the aforementioned report by HUMAN Security, traffic from malicious bots is also increasing.
It points out that the boundary between benevolent automation and malicious automation has become extremely blurred, and the question has shifted from 'bot or not' to 'trustworthy or not'.

In an era where AI transacts with AI, what is ultimately questioned is 'trust'.
Since AIs are responsible for matching, it is essential that the process is not a black box, actions are transparent, and monitoring is possible in a tamper-proof manner.

FANGO's answer to this is its patent-pending three-layer architecture.

[Patent Pending] FANGO 3-Layer Architecture

[Layer 1: Customer-side AI]
Operates in the user's local environment and holds the user's intentions and personal information.
Only the minimum necessary anonymized information is passed to the outside.

[Layer 2: Central Orchestration AI]
Stands between the two AIs, matches only anonymized conditions, and mediates negotiation and matching.

[Layer 3: Agent-side AI]
Manages information, including off-market properties, in the real estate company's local environment and discloses it only to parties that truly meet the conditions.

Furthermore, the action history and disclosure records of these AIs are recorded and monitored in a tamper-proof manner using blockchain.

What is recorded on the blockchain is not personal information itself, but only anonymized action trails.
Personal information remains on the user's local device. By sharing only the unalterable record of 'who accessed or presented what information and when,' impersonation and false information are blocked, while unauthorized use or leakage of information without the individual's consent can be traced and verified later.

In other words, blockchain is both a mechanism to eliminate malicious information and a mechanism to protect personal information and privacy.

The company has completed a patent application for this technology.
Patent Application 2026-033839: 'Real Estate Information System Using Multi-Layered AI Agents'

Release of the Beta Version of the Real Estate Consultation Portal for AI, 'fango.city'

The beta version that embodies the above concept is 'fango.city'.
It was released today as a demo.

Already, on the sides of lenders/borrowers and sellers/buyers, AIs that safely manage information in local environments and deeply understand the owner's intentions are operational.
fango.city is a portal for these AIs to exchange only the minimum necessary anonymized information and safely 'negotiate' with each other.

'fango.city' is what the company has envisioned since its establishment.