How should the telecommunications industry respond to the transformation that technology evolution – and especially AI – will drive? It’s a transformation that will affect strategy, services and operations, as Huawei’s Liu Kang explained at OTF2024.
The recent ninth Huawei Operations Transformation Forum (OTF2024) gave the attending global carriers, vertical industry leaders, and industry partners a lot of food for thought when they gathered in Istanbul on 31 October. The overarching theme of Huawei’s OTF2024 was All Intelligent Operations Enable New Growth, a theme that also informed the keynote speech of Liu Kang, President of Huawei ICT Marketing & Solution Sales.
With that theme in mind, he called his speech All Intelligent Operations towards the Future – but first took a brief look back to the launch of OTF in London in 2014. This was when, alongside global operators and partners, Huawei began what might be called the digital operations journey: exploring innovation and sharing best practices in the operations field.
Fast forward ten years and Huawei can boast showcases and positive business results for operator partners, boosting ARPU revenue and customer satisfaction and lowering churn. Awards, industry recognition and an expanding list of innovation topics have followed. All of which is encouraging, but no cause for Huawei – or anyone – to rest on its laurels. As Liu Kang noted, over the next decade, the evolution towards even greater intelligence in operations will accelerate, affecting interaction, personalization and telcos – or perhaps, given the changing requirements they face, they should be known as techcos.
As he pointed out, operators – and the industry as a whole – need to respond to the transformation technology evolution will drive – a transformation that will affect strategy, services and operations. This is because we are now firmly in the AI era. This is an era when end users want more device usage, higher speed, coverage everywhere and a greater variety of apps. Data traffic and volume are increasing. And interaction is becoming ever more diversified. Quite soon it won’t just be about simple touch. Multi-modal interaction will be key, driven by intelligent devices and AI agents.
Service modes will become more diversified too, affecting both people and industries. We’re living in a world where the total amount of intelligent devices will exceed 10 billion by 2030. What opportunities and challenges does this imply for our move towards all intelligent operations? For instance do we need to focus on AI for operations? Or operations for AI? The answer, Liu Kang suggested, is both.
AI for operations can make many things possible which were not possible before. Massive quantities of data are already usable. Much, much more will become available. This can be an asset; such data can enable intention interaction, cognition prediction, and intelligent analysis and decision-making. This is already enabling AI agents that can understand and respond to customer inquiries without human intervention – but we are getting closer to AI agents that not only perceive user experience and identify user intention but can use these to formulate precision campaigns in real time to provide truly personalized services.
What about operations for AI – the operations that must meet the experience requirements of multi-modal interaction in the AI era?
High-frequency AI interactions will be part of these experience requirements: an on-the-air interface latency of less than 20 milliseconds will be essential. There will also be a geometrical increase in operations and maintenance (O&M) complexity, caused by an increase in interaction efficiency, quality and object, that could be up to 100 times greater than it is now. And let’s not forget the likely transformation from network-oriented to service-oriented operations and O&M evolutions from ‘people and platform mode’ to ‘people and agent mode’.
Monetization
Bringing together both AI for operations and operations for AI is key, but there’s another challenge, as Liu Kang pointed out – the reshaping of four important business areas: monetization, user experience, O&M, and a concept that he called the digital life entry.
Let’s start with reshaping monetization. If we’re moving from a unified to a personalized offering, then the traffic monetization of today will give way to the experience monetization of tomorrow. To give an example, this could mean an operator using convergent data and AI to build user profiles, identify potential live streamers, and promote live streaming packages to them. Using this process, a successful 5G uplink monetization involving live streamers could have radical results, building user numbers, ARPU and revenue.
This process also involves information that tells service providers more about the user’s personality and habits; such data can be used to facilitate experience monetization. To whom? The markets are many and diverse: business travelers, gamers and takeaway riders are only three that could be successfully addressed.
Thus we have a future where intention identification and intelligent recommendation can allow us to ‘see’ and serve users' potential requirements. ‘4 Right’ is the favored term here: the right place, promoting the right offer to the right people through the right channel.
As Liu Kang explained, it’s about unleashing the value of targeted users and massive amounts of data, such as (to use a simple example) an operator that coordinates with coffee shops to build a real-time campaign based on location, thereby boosting sales.
User Experience
The second reshaping involves user experience. No longer is it about best effort. It’s about meeting the experience requirements of multi-modal interaction in AI era.
Here’s an example. Not long ago, Huawei and an operator partner jointly explored differentiated experience assurance for gaming and video services; an important part of this was, inevitably, end-to-end latency optimization. There are three stages to this process. First of all, Network Operations Center (NOC), Security Operations Center (SOC) and optimization collaboration needed to be incorporated into the system. Secondly, Huawei needed to streamline the data and process breakpoints among the maintenance, optimization and marketing with the help of a unified intelligent platform.
Finally a QCI revamp needs to be implemented: a different QoS class identifier (QCI) for different services. For example, QCI-3 is for WhatsApp and QCI-6 for VIP users. This is hierarchical management that can assure differentiated user experience. And it has already worked, helping a Huawei operator partner increase revenue by 10% and user satisfaction by 6.7%.
Looking into the future, Liu Kang explained that diversified services in the AI era require a focus, as we have already suggested, on the experience of latency and uplink. Huawei has built unique simulated reality of communication networks (SRCON) and digital twins to assure the best multi-modal interaction experience. This means that user experience can finally be perceived and guaranteed.
Reshaping O&M
The third part of the reshaping process we referred to earlier is reshaping the O&M mode, from traditional ‘people and platform mode’ to ‘people and agent mode’. AI is again key, of course.
For example, to accelerate the transformation from network-oriented to service-oriented O&M, a Huawei operator partner built an AI analysis engine, based on the massive data of routine experience, survey and, importantly, complaint. The idea was to identify and assess criticism. The sample data was enlarged 100 times, allowing for what is called detractor modeling. This could not be realized before.
Poor-quality cells were the problem. Through the collaboration between the NOC and SOC we referred to before, the poor-quality cells causing the complaints could be proactively rectified and the complaint ratio reduced – by 22% in this case. As Liu Kang made clear, this process needs to continue. In fact, in the future, the telecom foundation model and digital twins should be built in order to achieve ‘people and agent mode’, improving O&M efficiency and effectiveness.
Another example comes from a common operator pain point: FTTx maintenance. AI-based topology restoration can be performed to intelligently locate root causes of problems. The mean time to demarcation can thus be shortened and invalid work orders reduced. Meanwhile, FME Copilot reduces the amount of interactions between Feature Manipulation Engines (FMEs) and NOC by 50%. Put another way, AI-based O&M will be more proactive and efficient – with higher-quality outcomes.
Entering the Digital Life for End Users
With this future not too far off, it's clear that just being a communication services provider is not going to be enough; operators will have to being a digital intelligence life service providers. And that is the fourth reshaping: to reshape the way end users enter the digital life.
Some leading operators are building new entry points based on mobile money. Take, for example, an operator that signed up 300,000 international remittance users within five months, with an ARPU increase of US$5. Combining existing user relationships and data can be enabled with mobile money: for example, a 2 GB reward for account opening and transfer. A precision campaign like this can accelerate user growth.
But that’s just the start. The next step might be to build a SuperApp to provide diversified digital intelligence services, such as loans, insurance, taxis, or catering. And the SuperApp becomes an entry point into other digital intelligence services. Beyond mobile money different operators can build different entries to the digital life, based on their own specialties, such as Mobile Hall App, Cloud Phone, and New Calling.
One of the Chinese operators, for example, has just launched an AI assistant in its Mobile Hall App. Users can directly interact with the AI agent to get customized services such as traffic package, catering, and entertainment.
However, to make this work to its greatest advantage, business support system evolution needs to adjust, from cloud native to AI-native, in order to enable the intelligence of the entry point. AI is then utilized to improve call center functions, so average handling time can be reduced by 30%. Then massive customer service records can be analyzed for intention identification, bringing more business opportunities.
In other words AI-native business support systems are another part of the overall puzzle – one that will accelerate the development of digital intelligence services.
Or, as Liu Kang reminded listeners, the AI era has arrived. As his presentation showed, bringing together AI for operations and operations for AI will radically reshape concepts like monetization, user experience, O&M, and digital life entry. These changes will in turn help us to meet the requirements of the AI era, an exciting – but challenging – era of multi-modal interaction experience and accelerating digital intelligence services development.
