Recent advancements in AI are having a major impact on the telecommunications industry. In this exclusive digital interview sponsored by DXC Technology, we sit down with Ulrich Irnich, Chief Information Officer of Vodafone Germany to explore the implications on its operations, customers and networks.
AI isn’t necessarily new anymore. Vodafone, among others, have been using algorithms and machine learning throughout their networks and operations for years. However, it is the recent emergence of generative AI that is creating a new shift and dynamic for telcos to consider.
“Large language models mystically can generate a lot of sentences and words together. And this is fascinating for us humans, right?” Irnich queries.
The discussion turns to new opportunities, the CIO pointing to communication improvements and the ability to enhance the customer experience.
Critically, it is highlighted how technologies are enabling organizations to better meet changing consumer priorities. In the case of Vodafone, AI is enabling customers to view products through an increasingly detailed lens, reviewing the carbon footprints associated with tariffs or handsets to better inform purchasing decisions.
While generative AI is good for the customer, it is also good news for the employee, enabling them to work more efficiently and effectively – something the CIO describes as co-piloting between customer service agents and AI.
Listeners can also expect to hear how AI is helping Vodafone on the network side in relation to abuse management, cyber security and energy efficiency. Here, the CIO explains that by managing its virtual networks with AI, it can power down during phases of low activity such as at night to reduce energy consumption by 70% without impacting the CX.
To hear more about these topics, as well as the role of digital twin networks and importance of taking stock amidst technological hype, listen to the full interview.
Kevin Gray:
Hi, my name is Kevin Grey, publisher of Fierce Wireless and Fierce Telecom, and today I am here with Ulrich Irnich the CIO of Vodafone Germany. Ulrich, thanks so much for joining us today.
Ulrich Irnich:
Thank you for having me.
Kevin Gray:
So we're going to talk about everything from how AI and generative AI are impacting customer service to how it's impacting networks. But first things first, I hope that you could just give us more of a general understanding for how is AI really affecting today's business for telcos?
Ulrich Irnich:
I think we are at a super hype moment at the moment because everybody's talking about Open AI and ChatGPT because that's in all of our discussions. Looking back, I would say especially in telco, the artificial intelligence is already established since a decade and longer. So if you look to the different practices behind AI, you have a lot of machine learning stuff and deep learning and now on top there is a lot of activities going on on generative AI.
I give you some examples. Especially on machine learning, we have a lot of, let's say, recognition and pictures, and also on algorithms we are looking at our networks to understand and see where we have, for example, abuse or [inaudible 00:01:29] situations where we can optimize our networks. On the other side, everybody's very used to voice recognition. Especially if you're talking about machines or using chatbots, that's already a topic which is I would say pretty mature.
But I think we are now in the age of really automation mindset, I would say coming from the multichannel orientation where you have done a lot of automation and also AI stuff just for a specific department or an area. And now we are looking at more towards, I would say end-to-end processes and business processes. And here is a lot of opportunities also with gen AI when you're talking about that, especially large language models, when you are setting the baseline for that, and I would say the baseline is always looking at, I would say looking that the company is prepared to use gen AI.
You have seen that in the press that some of the companies released confidential data. That's one dimension. On the other side, when you're starting really generating new topics and also stuff behind that, you need also to make sure that you are fulfilling some guard rails which are more on the ethical side so that you make sure that you are not doing something with your company don't want to do. So that means you have to prepare an environment where people can really piloting and also using large language models and gen AI, not hurting the company but also experimenting on that.
Kevin Gray:
Great. So Ulrich, one of the things we were talking about before this particular interview was some of this stuff that is being thrown about in these conversations for AI. It's not necessarily new, right? People who have been doing chatbots for a while now. The question that I have for you is really what is new and how is generative AI really impacting what's already been done in AI?
Ulrich Irnich:
The new thing is, and that's the fascinating part especially if you look to ChatGPT or the things you are seeing from Open AI is that large language models mystically can generate a lot of sentences and words together. And this is fascinating for us humans because you're just typing in a keyword and you get generated a long text, which also gets a lot of people a bit scared, right? So what is then the humanity behind that? But let's look first on the opportunities such kind of technology will provide.
Now large language models, if you have really the environment for that, can generate text and communication, but it also can help you injecting knowledge from large, let's say, correspondence or discussions for you in short sentence. And I would give you an example, especially in customer service what we're using because especially, Kevin, what you said. Most of our customers starts the journey in chatbots and looking for digital companions in helping to solve a problem. But then at a certain point you come to a point where you say, okay, now I need to call someone because the chatbot cannot solve my issue.
And in reality then you start explaining the agent you are calling the same story what you did in the chatbot again. And here comes gen AI into play because if you log into your chatbot and did all your conversations, what we implemented now is that the agent sees in three sentence the summary of your half an hour chatbot conversation and can directly start where you stopped. And that's a big benefit for both, right? For the customer, right, because you do not need to repeat it again. And for the agent as well because he can directly focus on solving the issue of the customer, which is one of the good examples where gen AI can make a big difference as long as you have the environment to do that.
Kevin Gray:
I love how it actually has the ability to summarize it into that three sentences for that particular rep. One of the other examples that we were talking about that I thought is, again, super interesting actually is in more of the in-person element. So can you really describe how generative AI is helping out reps in the actual retail stores themselves?
Ulrich Irnich:
If you look to AI, we are pushing the next best activity for our customers and the best offer directly into their sales system. So that's number one. So if a customer visits our shops, the agent knows exactly what would be the best advice he can or she can give to our customer. Now besides that activity, he can also use co-piloting and entering let's say some keywords and get the proposition out of that system.
So that means if a customer wants to have let's say a best tariff, that's one dimension. If he wants now to add family to their pay plans and want to have here it comes now a very, let's say carbon footprint neutral way of tariff. So you as a customer can decide how much carbon you want to use when you purchase your mobile phone, your tariff plan, and also if you want to include your family including into that contract. And the sales rep gets that kind of information directly by entering some keywords.
Now looking a bit ahead, what we are also doing is going forward if you enter, let's say also our web shops, you will see your carbon footprint by each product you will select and if you're entering on the chat you want to lower that, the gen AI stuff directly, let's say suggested the best solution and also creates the product for you in the basket.
Kevin Gray:
Does it also help out with specific, I don't know, sentences, like forms of speech that the sales reps should actually be recommended to say as well or stay within or is it more on the product suggestion and recommendation level?
Ulrich Irnich:
It's mainly driven by the product. That's number one. And number two is now if you look to the automation mindset, we're also pushing, let's say co-piloting to the end of the, let's say the process. So on the sales rep and also on the agent to help them to automate their work as much as they can. And this gives a lot of benefit because they know exactly what is hindering them in their daily process in getting more efficient. So that's more, let's say employee effort, how much effort he has to play to utilize our internal tools and this kind of copilot helps them to reduce the effort. And nobody else knows that better than the ones who are doing the job every day, and so we are pushing that kind of thing really to them and helping them to automate their day-to-day work.
Kevin Gray:
That's awesome. I'm curious, are you calling it internally the co-pilot or does it have a name that you're giving it when you're giving it to the reps?
Ulrich Irnich:
No, at the moment we are calling that really co-piloting. There will be a different name, I'm pretty sure, but at that stage we are calling that co-pilot.
Kevin Gray:
Oh that's interesting. Okay, well great. So that's the customer service element and obviously there's a lot of innovation happening around there, which is really exciting. But for the remainder of this conversation I wanted to focus on networks. Similarly, we were talking about how AI has been in networks for a little bit already and it's already been applied to certain areas. So my question to you here is what's new here and how's generative AI impacting networks specifically on top of what's already been done?
Ulrich Irnich:
I would step a bit back and say, well what kind of AI is helping us already inside of the network? First of all, of course all this kind of abuse management. So that means if we see [inaudible 00:10:45] things happening in the network and one of our big proposition and also purpose is shaping the digital society. And that means also giving them the freedom to serve everywhere but being secure that they're not getting, let's say, compromised by I would say cybersecurity activities. So we are monitoring also the networks up to the router in the house of our customers, and if we see that there are compromise, we are fixing it and we are giving them as well some hints how he protects or she protects their home going forward. That's number one.
Number two is as we are looking to energy and also to our carbon footprint for our planet, we are managing our virtual networks with AI. So we are powering down on the consumption and the, let's say, movement of our customers in the night, the network and save 70% of our energy during that phase where there is low activities. And you can imagine when there are peaks of customers are coming into the network, we are powering immediately the system on, which helps us to give them a good customer experience and also helping us altogether in setting really something for our planet and reducing the carbon footprint.
Now looking at generative AI, the first basis is to having a digital twin. So to having really a digital copy of your physical network into a digital world, because this gives you everything possible, first on simulation. So where you want to invest going forward, your most. Network investments are really, really capital intense and we want to make sure that our investments are coming and also our customers are feeling that kind of investment. So that's number one what we are doing with AI. So that means on the digital twin we can simulate our capital investments immediately, how they affect going forward our, let's say our networks. That's number one.
And number two is if you have crisis. And we had two years ago a big flooding where we lost some of our mobile stations into that crisis and we used this digital twin and auto AI to identify where we have to place our mobile stations within 12 hours to make sure that connectivity is reestablished after 24 hours. And that's something which is also very important for our, I would say, digital society.
Now looking at gen AI, there are also other use cases, especially for the networks going forward. I would say we are still a bit in experimentation mode on that to be very honest. But we also want to make sure that we are using, let's say picture generating and also planning activities. One of the big activities we are doing is especially planning, so that means network deployment planning and all kinds of things. And we want to generate this kind of deployment activities with gen AI.
Kevin Gray:
That's super interesting. All right, so we have time for one more question here, Ulrich. So in particular, I think a good way to wrap this up, a lot of other telcos out there or even tech companies that are looking to implement new AI tech, they're really trying to figure out how can you define success. So if you really had to talk to any other CIOs that were out there or folks that were really looking to implement some of these new best practices, some of which we talked about today, how would you define success and what would that look like?
Ulrich Irnich:
First of all, especially in such kind of hype situation, the most important question I ask five times, why you want to implement? And if you have no good answer to that, just rethink it again, right? The key KPIs we are looking for is, of course, some of them are really productivity if you look to co-piloting. And people think, all right, then we don't need anyone anymore in our business, that's not the truth. Co-piloting and helping people to doing their job better and having less effort on that is one big driver, because customer satisfaction is driven by employee satisfaction. And if we are running down, let's say efforts people are spending in internal systems or what else, that's a big benefit. So that's number one. So productivity is a big thing.
Second is capability to be very honest. We are looking for hundreds of experts going forward in digital and IT, in tech environments. We will not find them to be very honest. So giving people capabilities and giving them opportunities in low-code, no-code solutions on co-piloting is a big benefit. This increase your innovation speed and it's increased as well everything where you have, for example, security topics where you need really experts you can utilize. So therefore if you have your why, please drill down how you measure your success and then mostly you will utilize that. But for me, besides productivity and financial impact, it's also for me on critical skills and how we can utilize technology to make us more capable. It's not to reduce us or to make us redundant, it's more how do we get more capabilities.
Kevin Gray:
All right, fascinating stuff. Well, Ulrich, that's all the time we have for today. Thank you so much for joining us and hopefully we'll be able to do this again some other time.
Ulrich Irnich:
Thank you very much, Kevin, for having me and thank you for all these very good questions. Thank you.