The "Deep Intelligence AI+ Real Estate Ecosystem Conference & 2026 New Product Launch" was held in Shanghai on May 27, 2026. Centered on the theme "Through the Door: The Modular Resonance of Real Estate," the event focused on the innovative application of AI technology across the broader real estate industry, including property development, senior living, housing rental, and property management services.
As one of the core products unveiled, the Deep Intelligence CoWork enterprise-level AI-native work platform was systematically showcased for the first time. Lyu Xinmiao, the product lead for Deep Intelligence CoWork, explained that while traditional AI tools can only handle basic Q&A and require manual secondary processing to integrate results, CoWork represents a disruptive upgrade. "I simply state the task objective, and it autonomously completes the planning and execution of the work. I only need to make choices among proposals and give the final approval, as it directly outputs a complete, deliverable work product. It truly integrates into internal enterprise work scenarios without requiring secondary assembly or processing," she stated.
In her view, CoWork's value extends beyond efficiency gains to breaking through traditional work boundaries. "Our goal this time is not merely an efficiency-boosting action. We hope that with it, individuals can independently complete more complex and challenging non-standardized tasks." This also signifies the product's evolution from the previous "question-and-answer" model to a formal "full-result autonomous delivery" model. The platform moves away from traditional functional displays, focusing instead on the entire real estate business workflow to intuitively demonstrate AI's core capability in reshaping industry work patterns.
On stage, Lyu Xinmiao demonstrated the platform's robust capabilities using the example of investment analysis for a land plot in Shanghai's Hongkou district. In the context of increasingly front-loaded project evaluation and intensified market competition, CoWork generated a complete feasibility study report covering 11 core modules—including market analysis, investment calculations, risk assessment, and bidding suggestions—in just five minutes. Unlike standard AI tools that require manual data compilation and condition setting, CoWork can autonomously connect to a company's local files, corporate land acquisition standards, the CRIC database, and the enterprise skill repository. It independently deconstructs tasks, integrates information, and produces tailored outcomes aligned with the company's specific standards.
Lyu Xinmiao highlighted the platform's human-AI collaboration and memory iteration capabilities. Addressing an initial deviation in unit type positioning for the land plot, she engaged in multiple rounds of human-AI interaction, adding requirements such as competitor benchmarking, land premium analysis, and differentiated competition strategies. The AI continuously iterated and optimized the proposal, ultimately outputting precise project positioning results. "It proposes solutions, and I make revisions. All work is completed within the same dialog interface. CoWork possesses memory capability, eliminating the need to repeatedly restate project background, thus ensuring work continuity," she explained. She emphasized that AI serves as a high-IQ professional assistant, with industry practitioners remaining the core decision-makers, and that human-AI collaboration will become the mainstream working model in the industry.
According to the introduction, CoWork fully integrates the entire business chain, including investment expansion, customer research, product development, pricing, and marketing. Leveraging its core Skills system, it can accumulate corporate operational standards, account styles, and expert experience. This enables capabilities such as refining customer profiles, generating unit layout proposals, conducting unit-specific price calculations, and producing multi-style marketing materials and videos in seconds, transforming individual professional expertise into reusable corporate digital assets. The platform currently comes pre-loaded with skills accumulated from nearly 80 industry experts, adapting to various non-standardized, complex tasks, truly understanding the industry, the enterprise, and the business. In the future, CoWork will continue to iterate and upgrade, assisting real estate enterprises in reconstructing work models, accumulating digital assets, and empowering the industry's intelligent transformation and upgrade.
Additionally, Zhou Xin, Chairman of the Board of E-House China and Chairman of CRIC Deep Intelligence, stated that CoWork is a genuine enterprise-level AI productivity foundation.
He pointed out that the industry's past efforts to build large language models internally incurred extremely high costs, often involving investments of hundreds of millions or even billions. Deep Intelligence offers a new co-construction model: "Our CoWork platform is an enterprise-level productivity tool for the AI era. My greatest hope is that with our industry large model as the foundation, all enterprises above it no longer need to spend billions or tens of billions building their own large models. On this basis, co-constructing an enterprise model with you is sufficient." Enterprises only need to build their exclusive second-layer enterprise model, enabling all employees to work efficiently based on the industry large model combined with their company-specific model.
Comments