KPN transforms its operations and empowers its workforce using Azure OpenAI

“From the moment we embarked on this technology, we knew that we needed to treat it like a big experimenting exercise. We said to ourselves: let’s try it, experiment it and see where it takes us.”

Jörg Kramer, Chief of Customer & Commercial Operations at KPN’s Digital, Retail and Customer Service division, is reflecting on his company’s Generative AI (GenAI) journey – and the cultural transformation this has triggered.

“Whenever I talk about GenAI, the first word that sprints to mind is always ‘game-changing’,” says Kramer. “That’s the beauty of this technology: it truly changes the game and how it’s played for any company. And we at KPN certainly don’t want to miss out on that.”

Being the largest IT and telecommunications provider in the Netherlands, KPN is also a leading adopter of GenAI in its industry and beyond. In collaboration with Microsoft, KPN has spent the past year embracing GenAI and using it to improve productivity, empower its employees and drive better services across its entire scope. All in a responsible and governance-centric way.

“At KPN, we go all out to connect everyone in the Netherlands to a sustainable future. We strive to enrich the lives of our customers using industry-leading telecommunications services and digital transformation,” he says. “Generative AI plays a huge role in helping us do it.

“It’s showing us patterns to modernize our operations and ways of working – pushing us to be more curious and find new and new ways to create value, every day.”

A global phenomenon sets innovation in motion

With its headquarter based in Rotterdam, KPN provides fixed and mobile networks for telephony, data and television to more than ten million people in The Netherlands. Founded nearly 150 years ago, the company currently employs just under 10,000 people.

“We may be a well-established brand, but we’re renowned innovators and tech advocates too,” says Kramer. “We strongly believe that in order to succeed in a sector as competitive as ours, you need to be a digital player and always stay ahead of the curve.”

That’s exactly what KPN has been up to in recent years. “We know our customers want a brand that’s easy, accessible and uncomplicated, so we’ve been concentrating our efforts into making services simpler and more digitalized for them,” he continues.

“The goal is to become an AI-powered organization that has human work at its very heart – and use this combination to lead the Netherlands into a more digital and sustainable future.”

AI has been playing an integral role to KPN’s digitalization efforts for years, but it wasn’t until 2023 that its more popular feature – GenAI – came into the equation.

“Our journey with GenAI started – similarly to many others – when ChatGPT came out, leaving the entire world hooked,” explains Winifred Andriessen, Director Advanced Analytics at KPN. “Of course, we at KPN were no different, and became interested in it right away.
“But what perhaps set us apart is that we didn’t just jump in on it for the sake of it: we went to Jörg [Kramer] and partnered up to identify possible use cases for his Retail and Customer Services department. We came back with not one, but five.

“It was the very beginning of our GenAI journey.”

Getting off on the right foot

Adopting new technology is never easy, particularly when it comes to artificial intelligence. From governance to skilling, every aspect needs to be carefully planned to make sure that AI is delivered in a responsible and ethical way.

Having had previous experience implementing AI for other parts of the business, KPN was well aware of these needs. So, when it came to implementing GenAI – and establishing a Generative AI Lab with it – the team already had a few ideas on where to start.

“The first thing we did was choosing the right provider for our AI vision,” explains Gianluigi Bardelloni, AI Lead at the KPN AI Excellence Center. “We had been already working with Microsoft Azure Speech Services in an Azure environment that we knew we could scale up quite quickly – so Azure OpenAI was a natural decision for us.”

Beyond their existing relationship, he adds, Microsoft proved the ideal partner because of the speed at which they made the implementation possible, as well as hands-on support provided coming from a range of committed consultants.

“Because we handle huge amounts of data – often also for the Dutch government – security is fundamental to what we do at KPN,” continues Andriessen. “So, when we have to select suppliers or partners, it's very important that we have good conversations about that.”

“That proved crucial when we decided to partner with Microsoft on this project. It really felt that we were helped and advised at all times – and at the same time we were operating in a completely secure environment.”

As Lienke Vet, AI Governance Lead at KPN, explains, the topic of responsible and ethical AI was another crucial point. “We are very much aware of the risks posed by Generative AI, so we wanted to make sure we use it in a responsible and contained way,” she says.
“We already had a responsible AI framework in place from our previous projects, but talking to Microsoft to define any additional measures was fundamental for us. In the end, Microsoft felt like they had an answer to all of our questions, but also a vision that felt trustworthy and allowed us to move fast.”

Getting into production

Once the due diligence was complete, it was time for the KPN team to start building the new environment and experiment with it.
Then, in January 2024, the first use case finally went live.

“Our first ever solution focuses on customer services and aims to make life easier for the team working there,” says Jörg Kramer. “This department handles thousands of calls a day, and its only visual record of them is a transcript that’s produced at their end automatically.

“To help cut down time reading through the transcript, we’ve developed a technology that creates detailed summaries of transcripts right after the call.”

Each summary produces four different types of information: why the call took place, what the agent did to support the customer, the outcome of the call and, if applicable, any scheduled follow-ups.

“That is revolutionary for us,” he says. “When you run a customer services operation, making the customer feel like you have their undevoted attention is crucial. GenAI allows us to do that and more – helping us to empower our agents to deliver the best possible services and make our customers happy.”

As next step for this solution, continues Bardelloni, GPT will generate additional takeaways tailored to the call center desk handling the call. We are also allowing our analysts to define and experiment with new takeaways for their desks,” he adds. 

“At the same time, we are setting up an automated pipeline which evaluates the quality of the generated summary - ultimately allowing us to intervene very quickly in case of performance deterioration.”

Building the foundation for a new future

The success of the first use case had led to fresh new opportunities at KPN – persuading the team to expand the solution and explore ways to reuse it in different settings, such as content writing.

“We want all of the copy and content we produce to be written in the tone of voice and following the brand guidelines,” explains Max Frinsel, Product Owner of Generative AI Lab at KPN. “To do that, we’re having GenAI write a draft based on them that it can then be edited and checked by the relevant team.”

There are also plans to work on future and existing digital assistants for both customer-facing and internal operations. “On the one hand, we’re working on enhancing our website’s digital assistant with answers it currently doesn’t know,” he says. “To do that, we’re equipping it with GenAI, which we’re training to generate one-time answers and seeing how they fare.

“At the same time, we’re developing a chatbot that our customer services agents and engineers can consult for support.”

But the most exciting part, concludes Kramer, is seeing people across the business get more and more involved with the technology and wanting to contribute to it.

"At the moment, 4,000 of our colleagues already have access to the call summary solution, and it's gaining momentum as more and more people realize the benefits it brings when it comes to providing better customer service,” he says.

“That’s what really matters to us. Despite some initial concerns, they’re starting to trust GenAI and seeing it as a truly useful ally rather than a threat to their jobs. A technology that’s safe, fast and empowering for us and for our customers.”