5G Factor: Telco GenAI Blooming

Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.

Transcript:

Ron Westfall: Hello, and welcome, everyone, to The 5G Factor. I’m Ron Westfall, Research Director here at The Futurum Group, and I’m joined here today, by my distinguished colleague Todd R Weiss, our Team Analyst focused on key areas such as 5G naturally and telecommunications. And today we will hone in on the major 5G ecosystem developments that have caught our eye. Well, Todd, welcome back to the 5G Factor and many thanks for joining today. How have you been bearing up between episodes?

Todd R Weiss: I’m doing great. Thanks again for having me. I love joining you on these, and now we’re right before Thanksgiving, so I’m very excited about Thanksgiving. It’s my favorite holiday. How about you?

Ron Westfall: Oh, I think it’s definitely a very good one and it’s definitely popular with many folks, and so it’s certainly welcomed. And so with that, let’s jump right in, and I think we’re tackling a very important hot topic, and that is telco generative AI-

Todd R Weiss: Like everything else, generative AI is going to be huge everywhere.

Ron Westfall: Well, yes, let alone AI and ML. Let’s not forget ML, and with that in mind, let’s start with Netcracker Technology debuting its GenAI telco solution a couple months ago at DTW23, and with that, they are looking to enable the communication service providers, which we commonly refer to as CSPs, to harness basically the transformative potential of generative AI to meet their specific use case demands. So, this is something I think is important that they’re breaking it out to address telco needs specifically. And so that is, I think, important because I think we’re going to see that across many industry verticals, and certainly telecommunications is at the top of the list.

Todd R Weiss: Absolutely, yeah.

Ron Westfall: Oh, yeah. Through Netcracker’s GenAI solution, CSPs can be positioned to, again, harness that telecom data and knowledge in a secure and controlled way to deliver business value gains to their customers and their partners, let alone their own businesses.

Todd R Weiss: Sure.

Ron Westfall: And to drill down more, the Netcracker GenAI solution includes support for telco-centric scenarios such as care assistant, agent partner, sale assistant, catalog assistant, and digital operations technician, and the solution builds on Netcracker’s well-established digital platform. And that is already extensively deployed throughout the ecosystem, certainly including the 5G ecosystem, and that includes top tier CSP networks such as NTT, Vodafone, Telefonica, Verizon, Deutsche Telekom or DT. You get the idea, it’s a who’s who of major telcos.

Todd R Weiss: It’s one of those things if you’re not on the list, you’re way behind.

Ron Westfall: Right on. I think that could be asserted, certainly. Yes, and I think the customer references are certainly built in. And so with that vast footprint, I think it’s a test to Netcracker’s ability to provide that innovation engine that I see as key to connecting the power of generative AI itself with the existing business support systems, operation support systems that are already deployed, and there’s just no way around that. Clearly, generative AI has to work with the existing BSS/OSS platforms out there. It’s not like they’re going to be ripped out and replaced because they want to adopt GenAI. That’s simply not going to happen, and so what I’m seeing is that through Netcracker’s established IT domain expertise in the telecom realm and the fact that the digital platform is already widely established, this can actually help drive adoption even accelerate of generative AI because it gives the CSPs the confidence that inner working with generative AI models can actually meet their distinct needs. And as a result, this can flow with intricate use cases that can be more developed, as well as the creation of high value services for both enterprise customers, as well as naturally consumer customers.

Todd R Weiss: Correct.

Ron Westfall: And so, what I think is a key takeaway here is that for the CSPs, most of that data functionality required to build generative AI use cases comes from their existing BSS/OSS solutions, where the Netcracker digital platform plays an integral role in being able to shore the credentials and capabilities of the new Netcracker GenAI telco solution. And so quite simply, it’s really providing that critical bridge of being a trusted platform to allow the operators to leverage existing crown jewels, the OSS/BSS platforms, and being able to also leverage emerging generative AI models to really be able to drive this innovation engine that is certainly being prioritized by the key decision makers in the CSP realm. And with that, I’ll turn it over to you, Todd. What about Netcracker’s new telco generative AI solution that leaps out at you?

Todd R Weiss: All right, Ron, from my view, a generative AI can be very transformative here throughout the telco domain, but only when it’s integrated with telco business support and operation support. That’s going to be a key, and I see one of the biggest challenges here across the telco industry is ensuring that generative AI is directly beneficial to the industry’s unique requirements. There’s all kinds of things, there’s networks, there’s bandwidth, there’s security, there’s a zillion things that this has to really work well with. The Netcracker GenAI telco solution is purpose built to directly fulfill the strategic goal of making GenAI highly valuable to the telco business. One of the examples they gave is significantly increasing productivity. Now, we see GenAI being used in things like bots and things like that in customer service, customer experience, things like that. It’s the same thing in the telco industry. There are a lot of customer facing things that they do. There are a lot of things that they can do where GenAI can make a difference, that can help them streamline their operations, have more capacity for doing things, things like that. More productivity is going to be important for the telco industry because they’re already a very busy business with a million things to have to deal with it.

Also, this Netcracker product also targets enriching the generative AI model with personalized prompts that provide all data, context, and instructions that can provide the best generative AI model results, which means they’re easier for their people to use because they’re being prompted, here’s how to do this. Great technology without knowledge is not an easy thing to use. It’s more difficult. So making it easier to use, making it more understandable, I think is going to be a very big part of a successful story here. And let’s see here, the telcos I don’t think will seek… Well, the telcos are trying to protect their sensitive data, like I said before, security is key. So, making sure that the GenAI helps protect sensitive data. They don’t want anything to be released to public generative AI models, they just want to maintain their own data. That’s a critical thing for enterprises and telcos of course. So, what Netcracker has done is they’ve embedded what they call ground-zero security to deliver this kind of security and safeguards, so that they can protect telcos. And I think the focus of this product on accurate and relevant results for telcos can foster the monetization of 5G cloud and edge investments, which companies always want to make that happen. They want to definitely have good arguments for doing this. They’re going to spend this money, there has to be real reasons to do this and there have to be giant benefits. So, I think this is something that will be very helpful to make those kinds of cases.

Ron Westfall: Yes, I think the motivation’s clearly there, and I think you made an excellent point, Todd. It’s about the business outcomes. It’s about ultimately monetizing their vast investment in 5G networks, and that’s something that I believe will happen. In fact, I anticipate that up to 90% of generative AI use cases that are driven by AI will require that BSS/OSS data foundation. That I think is just so fundamental that this is why we are seeing this type of solution being purpose developed, being marketed, and I think it’s definitely addressing not only a pain point for the operators, how do we monetize 5G better, but also driving a top priority or at least certainly meeting it, and I think that’s going to be a difference maker into certainly the foreseeable future because the telcos are not the only ones out there looking to how can we better use AI and ML technologies to improve our business and deliver better value to our customers. And so, great take, Todd, on the new Netcracker solution. And so, looking across the telco GenAI landscape, I see that Amdocs and NVIDIA are also collaborating on telco GenAI, really no surprise there, but I think what’s important is what is distinct about the partnership here. So, what they’re looking to do is customize basically enterprise-grade large language models that are running on NVIDIA accelerated computing as part of Amdocs’ amAIz framework, and I’m just going to call it amAIz. I think that could be a common reference for what the acronym is conveying is the fact that this framework can amaze. And so, what the collaboration is seeking to empower is that the CSPs can deploy generative AI use cases across their businesses and that includes very important capabilities such as network provisioning and improving customer experiences. And I think when you look at some of the details, they definitely, I think, convey why this partnership can drive more adoption of telco GenAI capabilities.

For instance, Amdocs will use NVIDIA’s DGX cloud, and that is an AI supercomputing platform, along with NVIDIA’s AI Enterprise software to again, support those adoption strategies as the CSPs are investigating how to safely use generative AI applications. I think that’s something that’s important to emphasize. We’ve seen with the early attempts with some of generative AI, that capabilities that produce things, what we call hallucinations. And so, I think what we’re learning is that yes, we want to of course avoid that, but we definitely want to have safeguards built in so that the telcos and any other major organization out there can have the confidence that not only is their data completely locked down from any threat, but also that it can be optimized for applying AI capabilities. And so alongside Amdocs’ strategy advancing generative AI use cases is the fact that the relationship with NVIDIA is already building on Amdocs’ existing relationship with Microsoft. So, now we’re seeing the train here gaining more momentum and Netcracker is doing the same alliance work with Microsoft. I think basically that’s requisite at this juncture in the market, that Microsoft is going to be a major player because of things like ChatGPT, and so forth, and driving telco generative AI use cases, but again, it’s an ecosystem play. It’s not just Microsoft, it’s NVIDIA and a host of others that we’re going to touch on here. And so with these new capabilities, including the NVIDIA NeMo framework, that is for custom LLM development and guardrail features that I just referred to.

So, this is really where the rubber hits the road. It’s enabling that the safety is built-in and that the large language model is distinct to the needs of the CSPs. And as a result, I can see that the CSPs will gain benefits such as improved performance, better resource utilization, and quite simply more flexible scaling as things like 5G standalone gets more built out. And also of key interest, the initial use cases are touching on things like customer care. We talked about that with the Netcracker solution, but it also includes resolution of customer inquiries by drawing more information from existing company data and it can certainly be dynamic, and on the network operations side, the partners are exploring how to proactively generate solutions that a configuration coverage or performance issues as they occur. And so, I think this is a sequel that is going to be much anticipated. This is where we’re going to see AI aiding automation to quite simply make operations more efficient, but also to align it with business outcomes in a more effective manner, et cetera. So yes, today it’s definitely benefiting things like customer care, customer experience outcomes, but also stay tuned, it’s also going to start impacting operations and business processes and ultimately how the entire businesses run. And with that, I know there are some other aspects here. Todd, what about the Amdocs/NVIDIA alliance that stands out for you?

Todd R Weiss: Well, first of all, just talking about NVIDIA, when you’re talking about anything with AI supercomputing, GPUs right away tells you that this is a trusted partner. Every customer is going to know it. Everyone’s going to know what they’re doing, and I’ve been covering this stuff for several years, their supercomputing stuff, their GPUs, all of these things from NVIDIA, the DGX cloud, the AI Enterprise software, these are hugely powerful tools. So, to work with Amdocs’ amAIz is going to be giant for the industry. So, that I think will right there, help take the telco industry into the place where they need to be, to be able to use these technologies. They’re not starting out with some little dinky partners. They’re starting out with the big guns, I think.

Ron Westfall: That’s really necessary, yes.

Todd R Weiss: Oh yeah, they have to. That’s credibility, as well as performance, knowing that they’re going in with a big partner like that, and that gives them credibility right away. So, I think that’s really important. So all this stuff, everything amAIz is doing with carrier, great architecture, open source, all these things, bringing in large language models, all of these things becomes a reality this way. And I think also through that alliance between Amdocs and NVIDIA, this better positions CSPs to provide services to their customers and to do things like… One of the things they’re going to do, automation of personalized assistants, all these kinds of generative AI tools that they’ll be able to use, and plus, it’ll be things like service ticket routing and use cases they haven’t even thought of yet. That’s the beauty of this, the innovation that can come out of this, I think, we don’t even know as far as what it can be yet because we’re just seeing the tips, the beginnings of it. It’s these big ideas that we’re starting with, but we’re going to see a whole ocean of things in the next couple of years, things we didn’t even think about where this could be used, I think. So, basically I think this is really exciting. I think that CSPs are looking for ways like this to find tools that let them do more with maybe even less, to do more to be able to have more customers, to do more to increase their capacity, to increase their services, to increase their bottom lines at the bottom of it all. So, I think this is really a big deal, and also one of the things that is also included here from this Amdocs product is metrics. So, they’ll be able to show quality of service metrics, things like that. All of these things will be measurable, so again, investments that are measurable, technology that’s measurable, that shows results. I think any time you spend money on an investment on a new product, a new technology, you need to be able to show that. So, I think that’s really important.

Ron Westfall: I agree. I think this is nascent, and I think what’s notable here is that both Netcracker and Amdocs are, I think, capturing early mind share. Now, yes, the telcos are still kicking the tires and generative AI capabilities are still very much at that beginning stage of, okay, how do we really leverage this? But it’s a journey, and so I think to your point, we definitely want to make sure that these generative AI models perform as required, i.e., the safety guardrails don’t break, for instance, but also it’s actually improving the value. It’s not something that is causing a distraction and so forth. And so, this is not necessarily unique to CSPs, but what’s refreshing is that these solutions are being developed specifically for the CSP requirements, which certainly requires that BSS/OSS data foundation. And I think that segues into our next topic, is the CSP themselves forming alliances to figure out, how do we best take advantage? How do we best utilize generative AI? And specifically I’m looking at South Korea Telecom, or SK Telecom, and Deutsche Telekom, or DT, have agreed to create a telco-specific large language model.

Todd R Weiss: It was only a matter of time. This was only a matter of time.

Ron Westfall: Yeah, certainly it’s already happened. It’s passed inevitability, and now it’s like, “Okay, this makes not only logical sense, but also it makes potential great business sense.”

Todd R Weiss: Oh, yeah.

Ron Westfall: This is for the operators and what I think is important to note here, it’s not exclusive to SK Telecom and DT. It’s really the first significant step that both of them are making since forming the Global Telco AI Alliance back in July, and that included e&, based in the United Arab Emirates, as well as Singtel. And what the pair have done is signed a letter of intent to jointly develop a multilingual LLM. So, I think that’s also going to be important. I think that’s certainly been something of a barrier. Obviously, having it English is the starting point, but what’s important here, it’s also going to be in German, and Korean, and so that will clearly facilitate adoption in those markets or certainly I believe it can make a difference with that. Then-

Todd R Weiss: And of course, we’re going to see others do the same thing and create their own telco models for something that they’re interested in. There’s not going to be just one telco large language model.

Ron Westfall: Oh, yes.

Todd R Weiss: There’s going to be a variety of these things.

Ron Westfall: Like e& using an Arab language one for example, Singtel and Chinese language for example. So, stay tuned and this is something that we’ll have… We can definitely anticipate more variation, and so specifically for SK Telecom and DT, they’re looking to have this ready by the first quarter of 2024. And what’s also important to note here, and we touched on this a little earlier, that this is completely an ecosystem effort and this is truly a journey. In fact, what they’ve done is that both companies have already established investments in the AI domain. For example, Anthropic, which scored $100 million investment from SKT back in August, as well as Meta, which in this area is known for its Llama 2 open source LLM, and it’s being distributed by Microsoft. So, they’re looking to co-develop it in partnership with these companies. And I think that’s definitely a smart play because it’s demonstrating that, “Okay, we have ecosystem flexibility, and this is something that can create even more partnership flexibility, as these models mature.” And as the LLMs are trained specifically for the telcos, I think it will create the opportunity to look at higher understanding of the telecommunication service related areas and the customer tensions. And as a result, this is something that is already, I think, having impact on customer service roles such as AI-enhanced contact centers and SKT is already implementing that, and we can anticipate that this will be even more impactful in Q1 ’24 and beyond as they start producing more tangible outcomes from the collaboration, and with that, what about this partnership is something that impresses you, Todd? What else is going on?

Todd R Weiss: Well, I think it’s a great model. This is going to be perhaps the start of what we’re going to see a lot of, I would imagine, in the telco business. Some of the businesses and industries were jumping on generative AI at the beginning. They could see where it could help them immediately. Telco took a little bit of time for it to start developing, and I think that’s what we’re seeing now. So with this model, this first alliance, this is a big deal. Certainly, there’s going to be others, and this large language model is intended to help telcos worldwide, Europe, Asia, Middle East to develop GenAI services, all kinds of things, but there’s going to be more development of these things because there’s going to be needs that specific industries or companies within Telco have that maybe can’t be addressed by this. But again, these large language models aim specifically at telcos. These are going to save time, they’re going to save money for these developing platforms as they’re brought together, and I think that it will give new business opportunities to telcos, to CSPs, to this whole industry, and growth engines for their businesses because it’s bringing these AI capabilities into this industry. And like I said before, customers yet don’t know what they can do with it. We’re still explaining what customers can do with these things, so that’s a big deal. One of the things I was impressed by, Ron, is DT, Deutsche Telekom has noted that a digital assistant based on a telco-specific large language model will be particularly good at understanding, summarizing, and answering these specific concerns, adding that it’s among the first beneficiaries from the joint development with SKT will be its Ask Magenta service bot, which is something they’ve been working on.

Ron Westfall: I think that’s an important point, that it’s already being tied to a specific product, a specific capability, and Ask Magenta service bot is, I think, something that can provide more incentive for the other CSPs across the 5G ecosystem to accelerate, not only their evaluation, but ultimately the adoption, the integration of generative AI to, again, achieve these benefits. And so we’ll definitely be revisiting this, this is something that is clearly a hot topic. 2023 is the year of AI, and we’ll see if the ’20s prove out to be the decade of AI as well. And it’s just interesting, for example, John Chambers observed that AI can have more impact as a technology trend than say cloud and 5G combined. So, you get the idea here that this is something that is very special and it has a potential to be, but we obviously have to be patient with the journey and figuring out how to, again, prevent the worst of scenarios in terms of how AI can make a difference and look at how it can really drive these business outcomes and deliver a great deal more value to partners, and of course, to customers and so forth. And so with that stay tuned vibe, thank you, Todd, for joining today on The 5G Factor.

Todd R Weiss: Thank you. It was a great pleasure to be here with you again, as usual.

Ron Westfall: Indeed, yes. So, I’m looking forward to our next conversation and to our listening and viewing audience, thank you again for joining us. Be sure to check out The 5G Factor at The Futurum Group website, as well as reserving our channel on social media, as well as YouTube and so forth. And with that, have a great 5G day, everyone.

Todd R Weiss: Thanks, you too. Take care, everyone.

Other insights from The Futurum Group:

Netcracker Debuts GenAI Telco Solution to Power Use Case Innovation

Amdocs and NVIDIA Ready to Power Telco Generative AI Use Cases

5G Factor VRN: Deutsche Telekom and Microsoft Azure Test New Private 5G Capabilities