Enterprise-grade AI goes far beyond simply answering questions: 3Chat AI drives customer conversations all the way to successful deal closure.
3Chat ai, a product of Shanghai Nuku Sanyan Technology Co., Ltd., has completed a 10 million RMB Pre-A round of financing in November 2026. This round of financing is jointly invested by MiraclePlus and strategic investors, with Lion Capital acting as the exclusive financial advisor. The funds will be used for large-scale customer acquisition through large model training for customer operation and team expansion for core positions.
Agent Runtime for long-cycle customer operation, turning customer conversations into actionable and closed-deal outcomes
Enterprises have accumulated a growing amount of customer conversation and behavior data, yet they still rely on employees to go through records and query CRM to judge who is worth following up, when to reach out, and what to recommend. Targeting enterprises with stable customer inquiries or customer assets whose transaction relies on multi-round communication, 3Chat ai provides a sales-result-oriented AI Agent system covering customer understanding, sales decision-making and business execution.
The system is supported by 3Chat ai's self-developed AgentRuntime for conversational marketing. Different from general-purpose Agents that complete single tasks, it is oriented to customers who require long-term operation: it continuously remembers the portrait, current stage, relationship status and key communication information on a per-customer basis; changes in customer messages, business data or agreed follow-up time points can trigger the Agent to re-judge the next step; the Agent calls corresponding knowledge, skills and tools according to the customer's current stage, and all proactive outreach actions go through unified authorization and business rule verification in advance.
Based on this operation mechanism, the 3Chat Agent is responsible for understanding customer needs and executing the next business action; 3Chat Builder allows business teams to configure customer portraits, stage strategies, as well as the knowledge, skills and tools available to the Agent in natural language, and the configuration can simulate multi-round conversations in the sandbox before going online; Evaluation combines real conversations and business results to identify strategy deviations, providing a basis for continuous optimization.
The Customer Growth solution for AI work assistants extends this customer operation capability to the daily work entry of employees. Through the MCP connector, 3Chat connects customer context, enterprise business rules and customer outreach capabilities to Doubao Work, WorkBuddy, Qianwen Office, as well as AI work assistants such as ChatGPT, Claude and Gemini. Employees can search for customers and formulate follow-up strategies in natural language, 3Chat screens crowds and pushes content according to customer tags and business rules, and then the Agent takes over customer replies and promotes subsequent communication.
The customer operation capability of 3Chat ai covers private domain scenarios at home and abroad, from domestic WeCom to overseas WhatsApp, supporting customer inquiry reception, demand identification and continuous follow-up. In the Xpeng Motors French market project, 3Chat undertakes customer communication after ad lead collection via WhatsApp, answers car purchase questions based on local vehicle model materials, identifies vehicle model preferences, car usage needs and test drive intentions in multi-round conversations, synchronizes customer portraits and communication summaries to CRM, and connects with local store sales teams. After the salesperson takes over, they can continue to use the previous communication information to avoid repeated inquiries, and promote the conversion of advertising leads to test drives and subsequent car purchase links. Project data shows that its operating cost has dropped from 144,000 euros to 70,000 euros, the cost of per-customer store visit has dropped from 250 euros to 148 euros, the first response time has been shortened from 28 hours to 30 seconds, and the in-store conversion rate has increased from 23% to 34%.
According to the company's existing product materials, the domestic version of 3Chat ai has served hundreds of enterprises, covering more than 20 social and e-commerce channels. Its customers involve retail, beauty, technology and other industries, and representative customers include Xpeng Motors, ByteDance, Estée Lauder, etc. Enterprises can select the product version according to the volume of customer reception and operation, and pay on an annual basis. In addition, 3Chat.ai has been successfully selected into Google for Startups Accelerator 2026, and continues to promote product market expansion and application for the global market.
To open the 3Chat customer operation large model Trio, invite partners to jointly expand industry applications
Trio is post-trained on the basis of open-source large models using authorized and desensitized real conversations between merchants and consumers, focusing on improving the shortcomings of general large models in sales scenarios: clarifying needs, advancing purchase decisions, handling objections, understanding long-term customer context, and making proactive recommendations at the right time. Its training and verification are all completed in the real business of 3Chat Agent. Along with the model, 3Chat ai will also release ConvertBench, an evaluation benchmark for conversational marketing conversion, to measure the model's performance on sales tasks in an open and reproducible manner.
For domestic CRM, CDP, SI manufacturers and channel partners, 3Chat ai is actively inviting partners to participate in product trials, scenario co-creation and cooperation exchanges. CRM manufacturers can embed Trio into their own products, SIs can use it to deliver sales Agents that pass acceptance faster, and channel partners can combine 3Chat's Agents and models to provide solutions for customers.
Team and CEO's Perspective
The team members come from universities including Harbin Institute of Technology, Tsinghua University and Shanghai Jiao Tong University, with rich experience in AI product R&D and many years of ToB marketing, sales and commercialization, covering AI models, product engineering, enterprise services, GTM and growth operation.
TAO Binjiang, Co-founder and CEO, has won honors such as Forbes 30 Under 30 and Fortune 40 Under 40, with many years of rich experience in enterprise service marketing; LI Tingting, CTO of AI Products, is a PhD from Tsinghua University, one of the former core leaders of the industry large model of Wenxin Yiyan, and has long been engaged in work related to large models.
CEO TAO Binjiang said: "We hope to turn every scattered customer conversation into a promotable, verifiable and closed-deal sales result. The projects with ByteDance, Estée Lauder and Xpeng Motors have convinced us that this set of capabilities can serve the most demanding customers. Next step, we will open up the precipitated marketing model Trio, so that more CRM manufacturers, system integrators and channel partners do not have to start from scratch, and can create deal-closing AI for their own customers."
In his view, the value of enterprise-level AI is extending from answering questions to understanding customers, executing processes and evaluating business results. 3Chat ai will continue to iterate products around this link, and work with partners to serve enterprises that need to promote conversion, repurchase and retention through customer conversations.