Laiye Technology: The Chinese Journey from Personal Assistant to RPA + Large Model Intelligent Automation

Origin

In 2015, Guanchun Wang founded Laiye Application, positioning it as a personal assistant service providing white-collar workers with personalized services such as scheduling, errands, and travel through human personnel assisted by large models. The model was heavy, had low gross margins, and was difficult to scale. Meanwhile, Rui Chu founded Oozon Tech, focusing on robotic process automation software, which had already been implemented in leading clients such as banks and insurance companies. The two routes merged into Laiye Technology in 2019, attempting to transform human assistant capabilities into enterprise-grade automation products and gradually moving toward a Chinese-style trajectory.

Milestones

Turning Points

• In 2019, Laiye Technology merged with Oozon Tech, leaping from manual assistants to enterprise-grade automation, marking the company's most critical business axis shift.
• In 2020, following financing, it upgraded from a tool-type product to an intelligent automation platform targeting large clients, successfully operating positive gross margin delivery for the first time.
• In 2023, it fully integrated large models, but encountered hallucinations and stability issues at client sites, forcing a fallback to a dual-layer design of rule-based systems supplemented by human intervention.

Failures & Pitfalls

• The Laiye Application personal assistant service failed to sustain scalability in less than a year due to heavy manual delivery and low gross margins.
• The assistant 'Xiaolai' saw user growth but lacked effective paid conversion, falling into a commercial predicament of having users without revenue.
• Early versions of the intelligent automation platform experienced large model hallucinations and stability issues at major client sites, causing some clients to revert to pure rule-based versions.

Core Data

• 合并后累计融资额：Hundreds of millions of yuan (including multiple rounds of strategic investment, based on public reports)
• 团队规模高峰：Peak team size of approximately 1,000+ people post-merger
• 客户覆盖行业：Large industries including banking, insurance, government affairs, telecom operators, and manufacturing
• 主要产品线：Robotic process automation software, intelligent automation platform, intelligent agent product matrix
• 核心商业模式：Software licenses plus implementation delivery, combined with agent execution volume-based pricing
• 2025年增长方向：Year-over-year growth in annual recurring revenue for financial customer service and ticket routing scenarios

Competitors / Peers

In the domestic automation and large model race, Laiye Technology primarily benchmarks against leading foreign industry peers, both of which started with robotic process automation and evolved toward intelligent agent platforms, with peak valuations once reaching tens of billions of dollars. Domestic direct competitors include HypercrAIt, UiPath-challenger Encoo, and Shadowbot. HypercrAIt is also a large enterprise automation vendor actively deploying intelligent agents in recent years; Encoo possesses strong delivery capabilities in finance and manufacturing; and Shadowbot enters through e-commerce and long-tail scenarios, acquiring customers with lightweight products. New competitors in the large model era also include open-source intelligent agent development platforms and agent-like workflow products built into large-model vendors, forming structural pressure on the intermediate layer of traditional automation vendors.