Businesses across the globe are dealing with an unprecedented amount of uncertainty. They must invest in capabilities, which make them adaptive, creative, and resilient. At Forrester, we call it being Future Fit. Being adaptive means having an ability to reconfigure your core business capabilities in response to changing market conditions to address customer needs. Being creative is about the ability to use data and insights effectively to create differentiating value propositions. Being resilient is about continuing to deliver on the brand promise, irrespective of the crisis.
These Future Fit imperatives apply across industries and organizations. Telecom industry itself has seen a significant transformation over the past few decades. Communications advancements have materially transformed the way we communicate and behave as a society. The move from voice to data and from cable to OTT are examples of how such shifts can disrupt the industry.
It would suffice to say that developing an emerging technology muscle is key to become adaptive, creative, and resilient. We will explore some of these emerging technologies here.
Generative AI is a newer advanced application of AI technologies that have come up in recent years. As the name suggests, Generative AI is capable of creating new data, images, videos, and even music. It can create entirely new content, based on the data it has been trained on, and it has a wide range of potential applications. Some of its capabilities are widely known because of the Open AI Foundation, which has created products such as ChatGPT (for text generation) and Dall-E (for image creation). There are many more such products in the industry, such as BERT, Google’s Bard, and Switch C for text and Stable Diffusion, Midjourney and more for images. By some estimates these technologies have received more than USD 15 bn of VC investments in 2022 alone with more to come.
While still in a somewhat nascent stages of development, these technologies have various implications on the telecom industry. One of the use cases would be to network performance improvement. One can, theoretically, use Generative AI to create simulated network traffic, which can help network operators optimize their networks for improved performance. This can help network operators identify and address performance issues before they impact the end-user experience.
Another potential application is in the creation of new services and applications. Generative AI can be used to create entirely new content, such as personalized music playlists, custom videos, and even virtual-reality experiences. This content can be used to create new services that are tailored to the needs and preferences of individual users.
Generative AI can also be used in the area of customer experience. For example, some of its text applications will become advanced enough to replace the automated chat bots that exist today. These AI applications are more advanced in holding a conversation. If one trains them on the organization support data, tickets, documents, then one can use them to create more organization-specific support tools. The potential of then augmenting it with other data and insights on customers with the firms’ products and services is immense. It can drive more personalized interactions with customers, leading to not just better service but also better cross-selling and up-selling of other services.
However, it is early days, and I would be remiss if I do not warn you about the potential risks associated with Generative AI. The biggest concerns, at least with the publicly available versions of these AI tools, which are trained using data from the internet, is the potential for bias in the algorithms. If the data used to train the algorithms is biased, the content generated by Generative AI is usually biased.
This often leads to negative consequences, such as discrimination or exclusion. Then there is this possible misuse through creating fake content. Bad actors are already using fake content to manipulate public opinion or spread disinformation. As Generative AI becomes more advanced, it will become increasingly difficult to distinguish between real and fake content, with significant implications to society.
Overall, it is a new and fast-emerging space with new innovations coming at a rapid pace. Telecommunications services providers must watch the evolution of these technologies to carefully consider its implications, and develop strategies to capitalize on them.
Turing bots are the AI that write code. It is hard to fathom this, but it is indeed one of the rapidly evolving areas within the overall Generative AI space. Imagine AI going through the design documents, architecture diagrams, and generating the code to create software, which delivers on these specifications. We already have solutions, which are doing this. In fact, one of the healthcare firms in Australia used these tools to generate 92 percent of the 180,000 lines of code it needed to write for its applications. Current versions of these tools are at an early stage. Even at this low level of maturity, the only thing you need to do is to provide a natural language prompt and the AI generates a code according to the instruction. For example – you want to have a form with five input fields labeled in a certain manner. You just have to prompt the AI to create that for you. You can further improve their design, positioning, format, simply using natural language commands.
At Forrester, we believe that in 2023 alone 10 percent of the worldwide code will be touched by TuringBots in one way or the other. What are these ways? The most straightforward application of this is in writing code as I illustrated in the previous paragraph. Other more likely use cases are in developing text scripts to test a certain piece of code. Or simply to read the code and create documentation around it. Such developments have significant implications on developer productivity.
Edge intelligence. Edge computing is an area that many in the industry already understand. Basically, it is about having computing power available at the extremities of the connected internet. However, this term has come into vogue in recent times due to the rising compute and processing power in the devices at the edge. This distributed computing paradigm enables the processing of data closer to the source, rather than transmitting the data to a central data center. This can help reduce latency and improve the performance of the network.
However, this is all in the past. The advanced problem in this space has moved further. We are looking at having AI-capable devices at the edge now. These devices can sense their surroundings and capture and process the data using pre-trained AI models deployed in them to generate local actions. The problem with this approach is how to make these pre-trained AI models learn from each other at the edge. As they operate in their local environment, over time, they become smarter in accurately sensing and understanding it. This smartness must be transferable to other devices on the edge. Currently, there is no elegant way of doing it. However, this is a growing field and there are firms that are working on this problem.
As a telecom provider, one must be aware of what this means for their current solutions for edge computing. As these solutions emerge, the amount of data that these devices carry back on the network backbone will come down and newer point-to-point needs might emerge. This will have implications on how you build these solutions.
In summary, I must say that emerging technologies are starting to emerge much more rapidly than before. It becomes very important for firms across industries to build in capabilities to understand and deploy them. However, keeping yourself up to date in such rapidly changing environment is not a trivial ask. They can also use various knowledge partners to make this happen. Not staying on top of these is not an option. These technologies are fundamentally changing the world and, in the process, completely killing existing business opportunities and creating new ones. To remain future-fit – learning about them, continuously experimenting and evaluating them to drive business outcomes is key.