CaiXun Co., Ltd. plans to acquire JiZhi Intelligent to expand its presence in the Voice AI Agent sector

On March 30, CAIXUN Co., Ltd. (300634) announced its plan to acquire 100% of the equity of Ji Zhi Intelligent, an enterprise-grade generative AI voice agent company. Based on publicly available information, this is one of the landmark acquisition deals in the A-share market this year in the enterprise Voice AI Agent (voice agent) sector.

After completing its “1+1+N” AI strategic layout, CAIXUN Co., Ltd. is attempting to further strengthen its core technologies and commercialization capabilities in the AI application layer through this outward/accretive acquisition.

Voice AI accelerates deployment

Over the past two years, the industry narrative around generative AI has mainly centered on the Copilot stage, which assists human workers. However, with the maturity of long-horizon agents such as Open Claw, AI applications are showing a trend toward evolution into “digital labor,” shifting from “assistive tools” to “directly delivering business outcomes.”

In high-frequency interaction scenarios such as smart marketing, intelligent customer service, government notifications, user activation, and insurance renewal, voice is not only the interaction method but also the core of business flow. Real-world business data shows that the introduction of Voice AI is significantly changing the input-output ratio in these fields. Taking financial and insurance scenarios as examples, leading Voice AI platforms have already achieved a clear improvement in reconnect rate and return on investment (ROI). Meanwhile, as the cost of underlying large-model inference continues to decline, large-scale private deployment of voice AI has for the first time become economically feasible in reality, prompting relevant vendors’ business models to shift from the traditional “SaaS model priced per seat/duration” to “pay per business outcome (RaS).”

Run the “generative dialogue outbound-calling” business loop

In this acquisition, the commercialization and go-to-market execution capability of the target company, Ji Zhi Intelligent, has become the focus of market attention.

From a technical architecture perspective, Ji Zhi Intelligent has achieved breakthroughs by deeply customizing vertical industry small models based on open-source models, and combining them with a mixture-of-experts architecture to reduce latency and improve realism. It also possesses engineering capabilities such as private deployment, streaming speech processing, and optimization of end-to-end Agent capability. Its actual business performance is already close to the level of human customer service, and it has completed the national generative AI service registration.

In terms of financial performance, Ji Zhi Intelligent has preliminarily run a commercialized business closed loop. According to a notice from CAIXUN Co., Ltd., Ji Zhi Intelligent achieved revenue of approximately RMB 130 million during January to August 2025, with net profit of about RMB 13 million (2024 revenue was RMB 150 million, and profit was about RMB 11 million). This performance is outstanding among similar start-up enterprises. At present, Ji Zhi Intelligent has built a customer base in vertical industries with high requirements for conversion rates, such as finance, automotive, and education, validating the business value of Voice AI Agents as “high-converting and measurable.”

For industrial acquisitions, the synergy effect after integration is a core consideration. Analysts point out that CAIXUN Co., Ltd. and Ji Zhi Intelligent have strong complementarity in their business layout.

First is complementarity in channels and customer segments. Ji Zhi Intelligent previously mainly served internet-based clients, facing challenges when expanding to large B-end enterprise users with complex tendering and bidding systems. CAIXUN Co., Ltd., having been deeply involved in major government and enterprise markets such as telecom operators, finance, energy, and transportation for more than 20 years, has a mature channel network. With Ji Zhi Intelligent being incorporated, it is expected that, with the comprehensive empowerment from the listed company, its business can accelerate scalable promotion.

Second is optimization of cost structure. CAIXUN Co., Ltd.’s reserves in voice channel resources and computing power infrastructure can help, to a certain extent, optimize Ji Zhi Intelligent’s cost structure, thereby improving its gross margin level.

Finally is the construction of a data ecosystem. One of the core barriers of Voice AI lies in real “execution-trajectory data.” By integrating both sides’ resources, it is expected to form a positive loop: “more real user interactions—greater volumes of high-quality Chinese speech data—better training of vertical models,” further strengthening industry barriers.

Computing power build-out and AI applications, driven by a dual engine

Judging from the performance of global capital markets, the Voice AI sector has been quite hot recently. After Sierra AI, a company in the overseas AI digital employee space, exceeded ARR (annual recurring revenue) of $100 million, its valuation reached the $10 billion range. ServiceNow’s acquisition of Moveworks also reflected a relatively high PS premium. By comparison, the transaction valuation of Ji Zhi Intelligent of about $50 million is within a relatively rational range.

It is worth noting that in the convertible bond issuance plan to unspecified investors released earlier this year, CAIXUN Co., Ltd. has already clearly disclosed its strategic plan to increase investment in “smart computing centers” and “AI applications.” This acquisition of Ji Zhi Intelligent can be seen as an substantive implementation of that strategic blueprint on the AI application layer.

According to performance commitments, Ji Zhi Intelligent’s cumulative net profit from 2026 to 2028 will not be less than RMB 84 million. This profit commitment provides a verifiable financial anchor for this acquisition, helping strengthen market expectations for its future growth.

Although the acquisition logic is clear, during the long-term implementation of strategy, the market still needs to objectively consider potential risks: first, integration risk—differences between the two parties in corporate culture and management models may cause the integration results to fall short of expectations; second, customer concentration risk—certain single customers of the target company contribute a relatively high proportion of revenue, creating some dependency; third, risk of intensified market competition—as major general-purpose large-model players expand their layout in the voice AI field, if they adopt aggressive pricing strategies, it may squeeze profit margins for vertical application vendors.

A massive amount of information, precise interpretation—available in the Sina Finance app