Core Insights - The market's attitude towards artificial intelligence (AI) has experienced significant fluctuations, moving from initial enthusiasm to a more cautious approach as businesses recognize the limitations of AI in making critical commercial decisions [3][4]. Group 1: AI Limitations and Business Needs - Current AI models primarily rely on publicly available internet data, which often contains unreliable information, leading to "hallucinations" where AI generates inaccurate responses [5]. - Businesses require concrete data for decision-making, such as repayment ability for loans and fulfillment records for procurement, rather than probabilistic predictions [5][6]. Group 2: Company Strategy and Business Model Shift - The company has shifted its business model from selling software (SaaS) to selling results (RaaS) to meet the evolving needs of enterprises that prefer definitive outcomes over analytical insights [9][10]. - A recent collaboration with Ant Group focuses on intelligent risk control, allowing banks to assess the authenticity of transactions and risk levels without extensive manual verification [11][12]. Group 3: Operational Efficiency and Cost Reduction - The company's systems have significantly reduced manual review costs by approximately 30% in financial reimbursement processes for manufacturing enterprises by automating the identification of fraudulent invoices [13]. Group 4: Future Plans and Data Infrastructure - The company aims to strengthen its data infrastructure in collaboration with the government, facilitating the flow of data across various sectors and ensuring data ownership clarity [15]. - The company is also addressing the challenges faced by Chinese enterprises in international markets, particularly in navigating complex tax and compliance requirements through its TaxSwift platform [16][18]. Group 5: Competitive Advantage - Future competition will hinge on the ability to manage high-value, accurate data rather than merely on the size of AI model parameters, positioning the company as a "calibrator" of large models using real data to provide truthful insights [19].

Core Insights - The market's attitude towards artificial intelligence (AI) has shifted dramatically by 2025, moving from initial enthusiasm to a more cautious approach due to AI's limitations in making serious business decisions [1] - AI models are primarily trained on publicly available internet data, which often contains unreliable information, leading to "hallucinations" where AI generates inaccurate or fabricated responses [1] - The company emphasizes the importance of providing "Ground Truth" in the AI era, focusing on verified data such as invoices and contracts that are legally recognized and immutable [2] Company Strategy - In 2025, the company shifted its business model from selling software (SaaS) to selling results (RaaS), responding to the complex environment businesses face and their desire for concrete outcomes rather than mere analysis [4] - A partnership with Ant Group was established to enhance intelligent risk control, allowing banks to assess the authenticity of transactions and risk levels without extensive manual verification [5] - The company's system has demonstrated tangible results, reducing manual review costs by approximately 30% in financial reimbursement processes for manufacturing enterprises [6] Future Plans - The company has clear plans for 2026, focusing on two main strategies: deepening its roots in data infrastructure and expanding outward to assist Chinese enterprises in navigating international markets [7] - As a technical support unit for the national public data resource registration platform, the company aims to connect disparate data sources across industries, ensuring data ownership and usability [8] - The company is developing the TaxSwift platform to help Chinese businesses comply with complex tax and regulatory requirements abroad, enhancing their global operational safety [9] - The future competition will hinge on the ability to manage high-value, accurate data rather than merely on the size of AI model parameters, positioning the company as a "calibrator" of large models using real data [9]