AI in All: Unveil Intelligent Landscape

Power in the Thought Man is but a reed, the feeblest thing in nature; but he is a thinking reed

• Pensees, Blaise Pascal

Even though man is as feeble as a reed, his power is endowed by nature with the ability to think. Intelligence is the unique characteristic of man that distinguishes him from any other species on this planet.

Now this unique ability is inherited by his creature: AI (Artificial Intelligence). The initial million years of hominid history produced tools to enhance, and later mechanize, mankind’s physical abilities to enable it to move faster, reach higher, and hit harder. Only in recent decades has computing brought the ultimate mental enhancement, especially AI which has shown the potential to realize any intellectual activity we can imagine.

AI is penetrating all sides of the IT system and radically changing how Telcos can solve problems, start new businesses, manage their organizations, and develop the market. AI has changed the ICT industry, and it will continue changing the industry more than any other invention of the past years, and it will come to pervade nearly all the relevant endeavors. We are just at the beginning of the AI revolution, however, today we try to offer just a glimpse of the potential impact of AI in the telecom application software.

As modern computing capabilities are becoming increasingly powerful in terms of storage and computing capacities, machine learning methods have gradually become dominant in the field of AI. Since 2022, LLM (Large Language Model) such as GPT (Generative Pre-trained Transformer) has swept through the realm of AI and has shown indications to surpass humans in many domains. This boom tide of AI, which is also the most profound in its over 60-year history since the famous Dartmouth Conference in 1956, is still ongoing and has shown no sign of ending yet. Although many essential problems are not yet resolved, AI waves have forever changed the ICT industry and application software.

Look through the Haze: Road to the Intelligence in Application Software

I know no way of judging the future but by the past. -- Patrick Henry You can never plan the future by the past. -- Edmund Burke

AI is not a single technology but a portfolio of technologies. CSP companies will possibly use these tools to varying degrees when applying AI in IT software, mainly BSS (Business Supporting System). Some will take an opportunistic approach, testing only one technology and piloting it in a specific function. Others may be bolder, adopting all and then absorbing them into their entire organization.

During its proliferation, AI has ignited the executives’ interest around the globe to whom many exciting things are possible with the newest evolution of this technology. Telecom executives are still trying to fathom the impact of how large language models and generative AI could exert on the application software competitiveness in the ICT industry. The AI use cases are especially prioritized in this hype. Among the telecom software juggernauts, for the main software products such as Billing, CRM, Mobile Money, Contact Center, and CaaS, a long list of potential AI scenarios is drafted, debated, scrutinized, and agreed to, but only to be identified as incomplete or outdated considering the rapid pace of innovation.

All Telcos have some certain ambition to experiment with AI and to take a preemptive advantage in the application software. However, every telco is facing the same imperative issue to ensure that adopting AI can bring more revenue and increase profitability and still guarantee the organization’s data, intellectual property, customer data security, and privacy protections.

For an early innovator, this requires a comprehensive AI scenario panorama map for a variety of telco leaders, from operations, the IT back office, the data team, and the business department, to take as the reference for implementing AI in application software. Hereby, AI could empower the telco instead of hampering or confusing the organization. At the same time, most early users of AI in telecom software have found it is critical to have distinct guidance of these AI scenarios that allow the working team to quickly understand and borrow when they meet the software issue in daily practice.

The impact of AI in application software is worthy of notice. Just as in the previous vicissitudes in ICT, the process of being mature and spreading of AI scenarios may not be linear, but it is very possible to move at an accelerating pace over time. In the furious tidings of AI-driven software practices, it is necessary to synthesize the trends and some conclusions could be deducted according to the observation and relevant analysis:

Dynamics of AI/Gen AI (generative AI):

• Compared with decision-making AI, generative AI brings disruptive changes to the industry. However, only the world's top ICT companies are capable of independently releasing disruptive software products.
• Gen AI will serve as the infrastructure to provide AI capabilities for various vertical software applications.
• In addition to the "superpowers" like Open AI, there are hundreds of emerging generative AI engines. However, the possibility of homogeneous competition among these AI players is greater.

Demands in the Market

• If the application software vendor owns overall technology and computing power advantages, it could apply generative AI in product-related scenarios to achieve leading competitiveness of products.
• The market space of the telecom industry is expected to expand profoundly due to the development of AI. New solutions, for either enterprise or consumer, are going to blossom in recent years. Myriads of untouched niche markets are envisaged and cultivated in various industries. AI provides a huge variable and possibilities for the industry competition landscape by instilling new technology in application software.
• AI engineering and openness are critical during the process of adopting AI. Telco is prone to manage the application software data, master the necessary skills, and exploit the software scenarios by itself.

AI Application Scenario Exploration

• The generative AI technologies apply to scenarios such as customer service chatbots, intelligent agent assistance, quality inspection, knowledge base, automatic generation of offering design, and marketing content. They can improve user experience, improve operation efficiency, and reduce the cost of building enterprise knowledge in a specific domain.
• Although there are some experiments on the business model related to AI for telco software like BSS, the example is MaaS (Model as a Service), it is still mostly acceptable that scenario-based intelligent solutions be sold.

Navigate with AI Cloud Map: Out of the Labyrinth of Pathfinding

I will liken him unto a wise man which built his house upon a rock. -- Matthew 07:24

When a new decision is made by the Board of Directors about telco application software with the spirit of entrepreneurship, the AI strategy should be part of its enterprise vision. This vision is expected to combine both the ambitious business target and the groundbreaking new technology in the software, from the frontend customer engagement tools of Contact Center and CRM to the backend systems of billing and digital wallet. This vision is always and must be built on a group of practicable AI principles which are illustrated in the “AI Practice Cube”.

The 1st dimension of the AI cube is the innovative and meanwhile feasible scenarios (use case). The user of telco can refer to them from scratch and probe the business domains which can be enhanced by adopting AI in application software.

Looking at the 2nd dimension, the accessible market is necessary whenever a telco wants to implement the technology. Since AI is an innovative force available everywhere, it will unlock new markets and lead the CSP to some niches that are not for the telecom solution previously. It could be compared with IoT or other revolutionary technology, but AI is regarded as bringing more fundamental changes to more domains in almost all industries.

The 3rd dimension of the cube is about the organization and enterprise culture to embrace AI. AI in application software will pervade every corner of the telco, just as the current IT tools are touchable everywhere. AI will not be a separate software, on the contrary, it is inside the backbone software to act as its gears and lubricant. It cannot and should not be treated as an optional software feature.

AI in All: Scenario Cloud Map

Since the 1st coordinate in the AI Practice Cube directs to the scenario, and almost all the ICT players hold enthusiasm towards the intelligence scenarios in application software, a scenario map is helpful to lead the decision makers and system designers to go through the maze of the AI experiment. Both the generative AI and decision-making AI users will benefit from it.

Facing the spreading wave of AI in almost all directions of telecom application software, a scenario map should be drafted for all the categories of business processes. This is the essence of the map – “AI in all”. Considering the understanding of AI scenarios in software varies obviously from the aspect of the users, many CIOs, IT managers, and business designers still feel confused even though they unanimously agree with the importance of AI. This could just be compared with the sailors on the unknown ocean who will always seek the map from others to have a reliable guide during their trip.

Pathfinders like Christopher Columbus and Ferdinand Magellan of AI in BSS frontier will take the map to verify the enormous possibilities and point to the worthy and valuable distant land. It could be concluded this full, distinct, and comprehensive picture of AI scenarios is a useful navigation tool when the application software team starts choosing and creating the AI feature on the existing or newly built IT platform.

This AI-in-all cloud-based map is exported from the industrial trends and underpinned by the IT system analysis.

• AI Trends in the Business Domains: there are three main telecom application software domains – Digital Communication, Digital life, and Digital Entertainment. Each domain requires some software product to solve or support the business processes. AI could be used to reduce manual operations and boost new sales.
• IT System Analysis: from the O&M side, the IT team works day in and day out to maintain the system and business supporting affairs. There is great potential to leverage AI to facilitate the maintenance and technical process.

Using the TMF-based Process Frameworx (eTOM) methodology, two business flows are used to organize the AI scenario for application software:

• Marketing to Lead to Cash: this is the process from the front end to the final revenue generation. The sales, marketing, and other processes related to customer attraction and conversion are in this part.
• Issue to Resolution to Cash: this flow helps the telco user to serve the existing customers, retain the customers, and promote more consumption.

Under this framework, the AI scenarios are divided into the categories of certain business applications:

• Lead Generation: the account managers and salesmen can adopt the AI to obtain better insight into the customer or prospective customer, and then they could identify the lead, and the opportunity, they can also evaluate the lead and arrange their activities with the assistance of AI aide.
• Intelligent Marketing: AI may be applied in the recommendation, generate the marketing content, or help the sales material collation. It could greatly save the effort of the marketing team and sales automation force department.
• Revenue Management: Since revenue growth strongly depends on the offering design and prevention of possible loss, it is critical to apply AI to the offering configuration and the future offering performance prediction, as well as fraud detection.
• Purchase and Promotion: when the application software can serve the customer with the virtual agent, it will relieve the numerous on-site and online staff. In this domain, the realization of AI will directly influence the end customer.
• Customer Service: there has been a lot of discussion in TM Forum about the potential AI scenarios in customer service, such as chatbots driven by generative AI. This application has been acknowledged as a main innovative direction.
• End-to-end Process Experience Improvement: customer experience analysis should not be neglected when a telco is building AI scenarios. The application software, such as customer engagement or contact center, tightly connects with the internal and external users. With AI, the experience index or personalization will soon be a critical necessity for customer satisfaction and retention.

A Tale in Heaven Country, No More Si-Fi

Chengdu city is known as the “heaven country” in China because of its comfortable weather and prosperity. In Chengdu City, Sichuan Province, China Mobile has built a call center using Huawei product AICC (AI contact center). The biggest challenge here is the agent efficiency. When answering a customer's call, the agent needs to record the customer's problem, search for related materials, and answer the customer's question. After the call ends, the agent needs to fill in the trouble ticket. This causes agents to be very busy and stressed, and prone to making mistakes or forgetting to follow up with customers. The solution is an intelligent agent assistant.

• When a customer makes a call, ASR (Automatic Speech Recognition) can be used to transcribe the voice call in real-time mode to avoid repeated questions.
• AI can intelligently recommend appropriate replies based on the questions described by customers.
• After reviewing the replies, agents can directly send them to customers to answer their questions.
• AI also listens to the agent's commitment to follow-up contact with the customer, generates a follow-up contact plan, and reminds the agent to ensure that every question of the customer is not lost.
• AI can also automatically summarize key information after a call ends, so the agents will not spend a long time filling in trouble tickets.
• The problems that customers often encounter are summarized and the chatbot provides answers to the customer to respond to his or her questions.

All of this dramatically reduces the time of customer issue resolution and relieves the burden of the agents. Both the agent and the customer benefit from the dexterity empowered by AI.

Finally, The AI-in-all cloud map is expected to provide a guide for most of the key scenarios in this direction. There is no need to elaborate on all the scenarios of AI in the telecom industry. It is particularly vital if this industry wants to unleash this groundbreaking technical power. It still requires massive amounts of data, data with high quality, structured or unstructured data, to feed the system so the silicon-based intelligence can improve its model and learn creatively.

The telco executive leaders need to plan now to build the scenarios for the application software so the IT system can raise the productivity level, and then the whole telco will benefit, from revenue growth to CapEx and OpEx reduction. Being contingent upon this condition, when a telco applies the AI-in-all scenario properly, AI can be a positive and practicable power instead of a pure Si-Fi phantom in the visible but untouchable mirage.