Delivering on innovative ideas and a passion for customer experience

Using Genesys Predictive Routing and a bold messaging architecture that encompasses WhatsApp and Apple Business Chat, Swisscom successfully reduced handling time and increased first-contact resolution while also improving NPS and agent satisfaction.

Watch the video

3% reduction

in average handling time


in NPS and agent satisfaction

20,000 monthly

customer enquiries in four languages handled via messaging

10% of all enquiries

solved via messaging without transfer to an agent

50% of

all FAQs successfully resolved by bots

5% uplift

in first-contact resolution

With Genesys Predictive Routing, we went from queue-based routing to strength-based routing around a specific KPI and customer type, all in a fraction of a second.

— Marcel Hischier, Product Manager Customer Interactions, Swisscom

Innovation partners

One of Switzerland’s most innovative and sustainable companies, Swisscom is the leading provider of mobile, network, internet and digital TV services for business and consumers.

Vital to its success are 4,000 sales and customer service agents who handle more than 50 million contacts annually, mostly incoming calls — in German, French, Italian and English. They also handle emails, chats and letters.

Those operations are powered by the on-premises Genesys™ solution, which also connects 16 contact centers with Genesys outbound, workload management, reporting, analytics and workforce engagement management tools. “We’ve been a long-term Genesys client for many years,” said Marcel Hischier, Product Manager Customer Interactions at Swisscom. “It’s more of a strategic relationship. We share the same passion for customer experience and regularly collaborate as innovation partners, coming up with ideas and experimenting.”

Building AI and machine learning into routing

Strong mutual interest lies in call routing; a subject Swisscom has turned into a science. The company uses it to not only reduce average handling time (AHT), but also ensure customers are connected the first time to agents with the right knowledge and skills. “Voice is our busiest channel, so the smallest improvements can make the biggest impact,” added Hischier. “Having built up experience of AI with our chatbots, we tested Genesys Predictive Routing to judge the future possibilities.”

Swisscom chose a 90-day fast-start pilot involving close to 1,000 agents in its German-speaking team. The pilot covered the first three stages of its four-stage approach:

  1. Discovery: Understand functions, benefits and metrics
  2. Predict and Prepare: Gather data
  3. Start Small and Get Ready: Select routing destinations and conduct A/B testing to see the effects
  4. Learn and Go Big: Expand use cases in short sprints using agile methodology

Offering services securely over hybrid cloud

Strictly adhering to data privacy laws, Genesys Professional Services helped feed data to the artificial intelligence (AI) engine using Swisscom performance information. All sensitive data was completely anonymized on site before it was sent to Genesys cloud-based AI services. Crucially, this meant algorithms were constantly updated and refined without the need for Swisscom to make any local system changes.

Through machine learning and algorithms, Swisscom better matched customer calls with the best-performing agents for different types of interactions. And, by switching between traditional and predictive routing, the company could accurately measure the effect. AHT was reduced by 3%. Net Promoter Score at specific touchpoints (TNPS) also improved slightly. There was no negative effect on other KPIs, such as the speed of answer and abandoned call levels.

“With Genesys Predictive Routing, we went from queue-based routing to strength-based routing around a specific KPI and customer type, all in a fraction of a second,” said Hischier. “We picked average handling time but could easily have chosen Net Promoter Score or sales orders. That’s what’s so exciting.”

Swisscom only needed minimal hardware. The pilot was easy to set up and it didn’t suffer from interruptions or downtime. Building on this success, Swisscom plans to reach its fourth phase, “Go Big” to develop more predictive routing use cases around other KPIs.

Our messaging bot resolved 10% of all inquiries, while the FAQ chatbot solved one in every two contacts. And we met our goal of achieving at least the same high customer satisfaction and sales conversion rates as our voice channel.

— Rolf Neukom, Product Manager, Emerging Channels and Bots, Swisscom

Blending messaging with bots and chat

Digital customer experience is another innovation focus area. Swisscom already has text, social and chat channels, and has added WhatsApp and Apple Business Chat as part of a blended messaging model.

“Although messaging is heralded as the new email, you’re still creating extra traffic and workloads to manage,” said Rolf Neukom, Product Manager, Emerging Channels and Bots at Swisscom. “Yet, we felt using messaging as a pure call deflection strategy was the wrong approach, especially if you offer premium-priced services. Instead, by blending it with other technologies, we removed complexity and enriched the customer experience.”

The Swisscom messaging architecture covers most reasons for contact, with the exception of contract cancellations, which a separate team handles. The architecture uses plug-ins between the company’s Genesys contact center solution and CRM and back-office systems.

In addition, Genesys Bot Gateways simplified the introduction of a fully integrated conversational chatbot service. It consists of concierge (for greeting, intent recognition and agent routing); FAQ (providing generic answers to the most common questions); and flow (solving the customers’ inquiries automatically with backend integrations).

Solving 10% of all inquiries and 50% of FAQs

Now fully live, the messaging solution enables 140 agents in seven locations to efficiently process approximately 20,000 customer inquiries a month.

Conversations are handled in four different languages; a chatbot offers asynchronous assistance. Topics range from mobile roaming and invoice queries to how to install broadband, WiFi and TV services.

Like many experiments, Swisscom achieved mixed results but met many of its original targets. “Our Messaging bot resolved 10% of all inquiries, while the FAQ chatbot solved one in every two contacts,” added Neukom. “And we met our goal of achieving at least the same high customer satisfaction and sales conversion rates as our voice channel.”

Importantly, by improving message quality, Swisscom also saw a 5% uplift in first-contact resolution. This led to a decrease in wait times, but it was more difficult to assess productivity savings with 90% of messages still requiring human intervention. Out of three possible entry points — IVR, SMS and web — most customers switched to WhatsApp or Apple Business Chat when browsing the company’s “Contact us” webpage.

“One of the big learnings was the need to step up from click-bot messaging to conversational experience, so agents could dig into the customer’s issue and remove ambiguity,” concluded Neukom. “With that adjustment, we plan to expand messaging to cover more sales and service use cases.”

At a glance

Customer: Swisscom

Industry: Telecom and IT service provider

Location: Switzerland

Company size: 19,000 employees with around 4,000 agents


  • Test predictive routing against specific KPIs and outcomes
  • Develop a blended messaging model