The first real estate AI agent "E-Home Xiaoxin" has launched full public beta testing.
On September 8, the world's first real estate AI agent "E-House Xiao Xin" was officially launched in Zhuhai, opening full public beta access for both property buyers and sellers. This means that Zhuhai residents can choose an AI agent to take charge of neutral intermediary services when purchasing or selling properties. E-House stated that during the public beta period ending on December 31, 2026, in addition to long-term access to 0-commission services from AI agents, a number of professional services will be completely free for a limited time.
The traditional real estate agency industry has long been plagued by practical problems such as commission disputes, information asymmetry, uneven service quality, and low transaction promotion efficiency. The root cause is that under the traditional model, the agency's income is directly linked to the transaction value of the property, which easily leads to behaviors such as creating information gaps and guiding rapid transactions. Consumers often need to bear high agency fees, and the fairness of transactions cannot be fully guaranteed.
Why can AI provide intermediary services? What kind of practical problems can AI solve? Yiren, CEO of E-House Xiao Xin, introduced that AI has advantages in three aspects: First, in terms of information opacity. AI Agent can retrieve and verify information based on full-volume data, which is not only from the public Internet, but also can perform various retrievals in professional databases, and give users timely responses at the first time. Therefore, AI Agent can retrieve information one by one from the housing source database, community database, supporting facility database, school district database, market condition database and policy database to give a response. Humans will forget to check things, but AI will not, so AI Agent has its own solution for information opacity. Second, in terms of high transaction costs. The intermediary work is done by AI, which does not charge commissions. The use of AI intermediary greatly reduces the cost of real estate transactions. Since AI has 0 commission, it has great advantages in the overall transaction cost. Third, in terms of low transaction efficiency. In the process of AI acting as an intermediary, AI can build very precise two-way portraits for buyers and sellers in various communication processes, then track the entire transaction process of both parties in the whole direct communication process, and promote the process. Therefore, AI-provided services can greatly improve the overall transaction efficiency. When humans track the entire transaction link, they will forget, omit various information and promotion steps, but AI can achieve full-process tracking.
AI is not an abstract concept. During the public beta of the E-House Xiao Xin product, users can directly consult E-House Xiao Xin on professional real estate issues, query information related to house purchase, and express their real property demand; they can either actively enter transaction demands through functional cards, or naturally express demands in dialogue and communication, so that AI can capture users' intention information. After receiving user demands, E-House Xiao Xin will match and output corresponding housing sources, and simultaneously push reference contents such as housing source details, transaction data, listing data, and market trends. When users are interested in a housing source, AI can create a group to facilitate direct communication between buyers and sellers; during the communication process of both parties, it will continuously answer various questions such as housing source details, professional policies, and market trends to promote the transaction. Follow-up actions such as booking house viewings, calling house companions, and entering contract preparation can all be triggered within the product. At each key node of the transaction, AI has set corresponding response actions to complete the full-link closed loop of intermediary matching.
As the core of the entire service system, the E-House Xiao Xin AI agent is no longer a simple auxiliary tool, but a transaction intelligent body that deeply participates in real estate transactions. This AI is built based on DeepLinkRE‑LLM, a large vertical model for the real estate industry. Relying on the four systems of database, knowledge base, expert base and engineering capability base, coupled with more than 20 years of real estate industry data accumulation from CRIC, it has built a complete business chain of "professional consultant - precise matching - smooth communication - service companion", with practical capabilities such as real housing source verification, property demand planning, intelligent matching of buying and selling demands, transaction process promotion, and intelligent answering of policies and regulations, which can directly complete business tasks instead of being limited to question and answer interaction.
Different from traditional human agents, real estate AI agents have four differentiated characteristics. The first is structural neutrality: AI has no interest demand for transaction completion, does not get income from commissions, has no motivation to hide property defects or release false housing sources, and will not favor any party to the transaction. The second is that the marginal cost is close to zero: the cost difference between serving a single customer and serving a large number of customers is small, enabling high-value intermediary intellectual services to achieve marginal free access. The third is that the boundary of information reserve is wider: it can integrate massive data of housing sources, market conditions, policies and regulations, without being limited by individual experience and service radius. The fourth is 7×24-hour continuous online access: it can quickly complete massive information retrieval, comparative analysis and supply-demand matching, and stably output intermediary services. Relying on the above characteristics, the AI agent realizes 0 commission for the whole process of intermediary services, eliminating the motivation of interest bias from the root of the mechanism.
Real estate transactions belong to the disposal of large-value assets, with long transaction chains and many offline practical nodes. Relying solely on AI cannot cover all real scenarios. E-House Xiao Xin does not choose to completely replace agent positions with AI, but sets up the role of "house companion" to supplement the value of human beings. As a dedicated contact person, the house companion will follow up all nodes of the entire transaction process from the verification and optimization of housing sources, and call AI capabilities or mobilize various professional service resources according to actual scenarios to ensure that both buyers and sellers always have a dedicated person to contact, and there will be no breakpoints in the transaction process. AI is responsible for efficiently processing massive information and completing supply-demand matching, while the house companion is responsible for interpersonal communication and process control. The two jointly undertake the intermediary companion function of traditional agents, and the role itself does not participate in commission sharing, and is not linked to the transaction result in terms of interests.
In the offline practical link, an independent professional performance team will take charge of the work, forming the third layer of service support. Businesses that highly rely on offline operation and professional qualifications, such as pre-house inspection, contract signing, loan and certificate processing, and lawyer witness, are separated from AI intermediary and house companion services, with clear price tags, optional services on demand, and billing based on workload.
It is worth noting that in this public beta release, E-House Xiao Xin also announced transaction certainty guarantee services provided in cooperation with professional institutions: full-process lawyer witness and companion, covering 5 major legal guarantees; 108 pre-house inspection services, covering various inspection points to check property risks in advance; the signing service center is also open simultaneously, which uniformly undertakes the implementation of offline signing, loan and certificate processing, transfer agency and other services, ensuring that the process has a dedicated venue and the services have a responsible entity. The above professional services are currently undertaken by professional partners such as Duan & Duan Law Firm and Xiangguchui House Inspection Institution.
Looking back at the product landing process, E-House Xiao Xin completed its first release on May 27, explaining the AI zero-commission transaction mode to the public for the first time; in August, it launched the housing source side public beta in Zhuhai in cooperation with local real estate enterprises such as Huafa, accumulating real housing source supply. By the time the full public beta was launched, the platform had completed the upload, verification and online release of more than 500 housing sources, and accessed Huafa's new property projects at the same time, with its service scope covering both second-hand property and new property purchase demands. This full public beta for both buyers and sellers marks that the platform has ended the single housing source side test and officially run through the complete closed loop of buying and selling transactions.
For this Zhuhai pilot, industry insiders believe that commission conflicts and information opacity are long-standing consumer pain points in the real estate agency industry. E-House Xiao Xin's pilot of the new mode of "AI neutral intermediary, dedicated person companion connection, and professional institution performance" in Zhuhai is a meaningful innovative attempt. It is expected that the pilot can precipitate replicable and promotable industry experience, promote the industry to achieve standardized development with the help of AI technology, and let the technological dividends benefit ordinary consumers.