From Efficiency to Resilience: Strengthening Enterprise AI Applications

        In 2022, generative AI will emerge, not only significantly lowering the threshold for using AI, but also lowering the threshold for using all digital tools. After two to three years, this AI craze has become ubiquitous, but it has also caused anxiety among some business executives: "How can we start using AI

        When enterprises import generative AI, there are four important steps:--

                Identify opportunities: The key areas that AI can impact include automation, supply chain management, product innovation, and customer interaction. AI is not about creating highlights, but rather addressing pain points and avoiding 'AI for AI's sake'.

        Choosing tools and algorithms: Training large language models is not something that enterprises need to do. The AI hardware equipment provider is Huida, whose clients are international platform companies such as AWS, Meta, and Microsoft. The platforms and computing power they build are intended for use by enterprises. "If you want to drink milk, you don't have to raise your own cows." Enterprises need to find a suitable farm for themselves. Choose appropriate AI, choose tools based on the company's operational needs, and integrate them with the existing operating environment to reduce the organization's learning costs.

        Preparation of materials: Inventory the "available materials", upload them to the cloud and make them accessible. Through AI, it is possible to break away from traditional methods and break through the complex steps of data silos.

        Execution import: Not limited to the technical department, ensure that employees at all levels can participate, and senior managers should lead by example, making AI a common tool for the entire organization, thereby improving work quality, continuously innovating, and balancing work and family.

        Provide reasons for change

        According to the Microsoft Work Trend Index survey, 70% of employees are willing to use AI to reduce their workload; But 49% of employees are also concerned that their jobs will be replaced by AI. Microsoft CEO Satya Nadella was asked at the 2023 CEO Summit how to encourage employees to use tools that may replace them in the future? Satya answered, trying to provide reasons for change and make them believe.

        Not teaching the people how to fight is like abandoning it. "Empowering colleagues through AI requires strategic and sequential promotion and training, trial and reward, promotion and internalization, so that colleagues are willing to learn and embrace AI tools. Currently, Far EasTone Telecommunication has 150 AI applications internally, saving 5143 minutes per month. For enterprises, the ultimate goal of introducing AI is to reduce costs and increase revenue. The former improves productivity and operational efficiency of processes, systems, and organizations, while the latter involves new products, services, features, and business models, and enhances customer satisfaction.

        Far EasTone Telecommunication's AI application has proven successful

        Far EasTone Telecommunication widely applies M365 Copilot in the daily work of employees, encouraging various departments to find application scenarios for generative AI, including product management, retail, legal, accounting and administrative facilities, procurement, IT development and other units, significantly replacing repetitive human work. For example, using Copilot to review leases and analyze summary files reduces average working hours by 66% and 95%, respectively; Automated document login, repair management, review and maintenance contracts, saving over 2000 minutes of work hours per month; Verify suppliers and manage expired contracts, significantly reducing working hours by 75% to 90%; Anti fraud data verification KYC authentication mechanism reduces 60% to 99% of working hours.

        Far EasTone Telecommunications IT Operations Department uses generative AI to optimize cross team operations work, allowing experts to collaborate with the "Operations Intelligence" AI assistant to achieve self-learning and innovative applications, quickly replicate application scenarios, share operations platforms, optimize operations efficiency, and reduce manpower for analyzing events by 30% to 50%. The network technology team also utilizes AI to integrate past accident records, over 300 operation process manuals and documents, quickly locate problem areas, automate order scheduling, accelerate learning and troubleshooting, issue early risk warnings, and enhance network operational efficiency and resilience.

        Moreover, Far EasTone Telecommunication will incorporate generative AI into its products and solutions to enhance customer experience and expand new business opportunities. For example, friDay Video uses generative AI to create conversational recommended playlists, combined with multiple technologies such as semantic and emotional analysis, keyword search, and semantic vector search, resulting in a 7-fold increase in click through rates; With the assistance of AI, the sales performance of Double 11 increased by 1.6 times compared to last year, and the number of newly registered members also increased by more than 2 times.

        On the enterprise side, Far EasTone Telecommunication integrates AI into its self-developed Internet of Things (IoT) platform, supporting multiple Yuan Ze University smart city scenarios. AI is used to optimize electricity demand forecasting, equipment energy-saving control, street lamp monitoring and regulation, and combined with Digital Twin technology to create a virtual real integration platform that can instantly link, simulate the future, and predict risks, improving decision-making efficiency. In addition, the self-developed 'Far EasTone Telecommunication Intelligence Platform' has been utilized to introduce generative AI to enterprises, with achievements spanning across industries such as government, schools, hospitals, technology, finance, ports, transportation, and food retail.

        At the same time, Far EasTone Telecommunication also utilized Copilot, combined with the "zero error" knowledge of world-class crisis management expert Dr. Qiu Qiang, to identify one-way weaknesses in the new product development process documents through rule-based error rate calculation rules and knowledge-based error rate calculation rules, and used Copilot to identify omissions in the documents under review, deepening the "zero error" culture within the company.

                AI is not a flash in the pan

        In the future, AI will move towards the "agent model". To keep up with the AI transformation, enterprises not only need to re-examine their organization and operational processes, but also use AI to build the enterprise brain and talent inheritance, drive safe operations, and continuously innovate and grow. Only by amplifying AI synergies can we create multidimensional enterprise resilience and achieve sustainable operation. "This time, AI is not a flash in the pan! (This article is a compilation of excerpts from Chee Ching President's speech and keynote speech at the Economic Daily 2025 Innovation Forum.)

        #