Recently, the AI application startup Lingxi Technology has successfully completed the closed loop from technology to performance in high-barrier sales sectors such as insurance and finance through its self-developed causal large model. It announced that it will achieve scaled profitability and positive cash flow by 2025, providing a new paradigm for the practical application of large models in the industry.
For a long time, B-end enterprises have had relatively low sense of achievement regarding the application of large models. Although general large models perform well in dialogue understanding, they often suffer from "hallucinations" and unstable decision-making when facing highly logical and professional sales scenarios such as insurance and finance, making it difficult to directly replace humans in complex sales tasks. Many companies face an awkward situation of "high investment, difficult delivery, and unstable results," with large models in business frontlines often seen as tools that can only chat but cannot generate actual performance.
To address this pain point, Lingxi Technology did not follow the mainstream trend of competing on model scale but focused on enhancing the AI's "attribution" and decision-making capabilities. Its core technical solution focuses on "causal AI" and "post-training of large models." By transforming the business logic of sales experts into causal judgment benchmarks, AI is no longer mechanically repeating scripts but can understand customers' underlying unspoken meanings like top salespeople and re-examine decision paths in real time. This "post-training" mechanism enables the model to not only absorb industry knowledge but also learn complex business decision-making logic.
In terms of business model, the company has implemented a "Results as a Service" (RaaS) model, deeply aligning its interests with clients' business growth. Unlike the traditional SaaS system's "pay first, bet on results" logic, this model quantifies the value of AI based on key business indicators such as premium increases and revenue growth. According to data, after insurance clients connected to its sales agent, new premiums reached 2 billion yuan within one year. As the model's autonomous decision-making capabilities improve, the task replacement rate of AI in vertical industries has gradually increased from 30% in the early stages to nearly full process autonomy.
Currently, this solution based on Customer Engagement Agent (ACE) has been scaled and applied in multiple fields such as automotive, banking, and education, with partners including industry-leading companies like Chery and Gaochu. The core team of Lingxi Technology mainly comes from Baidu's artificial intelligence department, with more than ten years of practical AI experience.
Industry analysts believe that the ultimate value of AI lies not in parameter competition, but in whether it can create incremental value in the physical industry. The case of Lingxi Technology proves that when AI evolves from a "tool" to a "productivity" and directly delivers business results, the commercialization dilemma of large models will see substantial breakthroughs. With the explosive growth of large model applications in China, leading enterprises with result delivery capabilities are beginning to realize their value first.
