Generative AI technology has swept the globe since last year, not only shining brightly in the work and study of ordinary people, but also playing an important role in the enterprise sector. Taiwan Mobile Enterprise Services, in collaboration with Taiwan Cloud, successfully upgraded its intelligent customer service system, Xiaomai, using the "Formosa Foundation Model" and AFS (AI Foundry Service), bringing a brand-new improvement to customer experience and employee productivity.
In close collaboration with Taizhi Cloud, Taiwan Mobile's technical advisory team utilized AI 2.0 tools and technologies. Through optimized parameter adjustments, embedded models, and a customer service knowledge base, they successfully created a customized language model for enterprise users. This made the intelligent customer service assistant, Xiaomai, more accurate, reducing the time pressure on customer service staff when dealing with customer inquiries, improving customer service team retention, enhancing manpower stability, and improving real-time customer experience and customer satisfaction.
Integrating service capabilities and experience to comprehensively meet the implementation needs of enterprises.
Taiwan Mobile's technical advisory team not only upgraded the intelligent customer service system, Xiaomai, with AI 2.0, but also accumulated rich practical experience, including raw data conversion, accurate corpus input, output of expected results, and overcoming bottlenecks encountered in model training and optimization. Taiwan Mobile has over 20 years of practical experience serving enterprise transformation across manufacturing, retail, healthcare, and finance sectors. This year, it will combine Formosat's model and AI tools with Taiwan Mobile's AI solution to launch the "AI 2.0 Solution," helping enterprises tailor their own generative AI. This solution leverages the Formosat model's deployability and Traditional Chinese capabilities, combined with Taiwan Mobile's end-to-end one-stop service, to accelerate the adoption of generative AI applications.
Wang Baoqing, Vice President of Enterprise Services at Taiwan Mobile, pointed out that in the AI 1.0 era, the entry barrier for AI technology was quite high, limiting its use to large-scale enterprises with ample IT resources. Besides collecting large amounts of data in the early stages, data screening and labeling by annotators, as well as hiring data scientists and AI engineers, were required for AI model training that could last for several months before POC (Proof-of-Concept) verification could be conducted in a real-world environment. The overall AI project execution time was as long as 18-24 months, and the subsequent results were often unpredictable, leading many enterprises to lose interest in promoting related projects. In the AI 2.0 era, all of the aforementioned problems can be easily solved with the assistance of Taiwan Mobile's professional technical consulting team. Enterprises only need to introduce the Formosa large model from Taiwan Mobile's AI platform, combined with their existing domain data, to complete model training and deployment verification in about 1-2 weeks. This not only saves considerable time and costs but also allows them to easily leverage the various benefits of generative AI. AI technology can now be applied to SMEs and traditional industries, which is a very important direction for Taiwan Mobile's development of generative AI services.
Exploring the infinite possibilities of generative AI applications for enterprises
When it comes to generative AI applications in enterprises, many people immediately think of intelligent customer service, assuming that only B2C businesses have this need. However, this is not the case. Chen Fangzhi, Deputy Director of the Information Systems Architecture Department at Taiwan Mobile, suggests that companies can leverage generative AI to optimize workflows from the perspective of internal customer service. For example, every year, companies have new employees joining. In the past, HR personnel would guide these new employees through the onboarding process one-on-one. Now, generative AI can be used to create HR robots that guide new employees through the onboarding process independently and answer various questions, helping them better understand the company. On the other hand, generative AI can also be integrated with knowledge management systems, allowing employees to extract information through inquiries, saving time spent searching and integrating data themselves. For instance, there are many reasons why manufacturing machines need maintenance. When new engineers encounter unfamiliar situations, they may need to consult senior colleagues or search for information and answers in the knowledge management system. In contrast, with generative AI, engineers can simply ask the AI for guidance.
To meet the diverse needs of various industries, Taiwan Mobile will leverage its cloud and 5G resources, along with its experience serving vertical industries, to become a dedicated AI consultant for businesses. This solution will assist clients in resolving various challenges encountered when implementing generative AI. The "AI 2.0 Solution" allows enterprise users to complete model training and validation within 1-2 weeks. This not only saves time and costs but also fully utilizes the various benefits of generative AI, promoting the application of AI technology across industries, facilitating digital transformation, and helping businesses flexibly identify new, high-efficiency, and low-risk development opportunities.