Does Automation Make Customer Service Smart?

Sebastian Glock, Cognigy

Poor customer service and perseverance in long waiting loops – what we have begrudgingly accepted as the norm can be changed with the help of modern AI software.

Cognigy, an internationally successful pioneer in customer service automation, offers one of the leading Conversational AI platforms (think voice- and chatbots). Companies such as Bosch, Daimler, Henkel, Lufthansa and, more recently, BioNTech are automating their customer communication in a smart way.

We spoke with Sebastian Glock, Technology Evangelist of Cognigy, who explains why bots have an image problem and how companies can actually save on service costs while making customers happier.

As customers, we always need individual support from companies, whether it’s questions about products, contracts, changes or problems. Chatbots rarely help. Do you see it that way?

Glock: Definitely! 90% of FAQ bots are unfortunately useless – even those on the websites of big brands. As customers, chatbots often make us feel blocked from getting the information we are looking for. The result? We call the service center more annoyed. For most of us, FAQ chatbots do not solve our problems, do not save any costs in the call or contact center and weakens rather than strengthens customer loyalty.

Why are chatbots used in the first place?

Glock: With thoughtful applications and smart implementation, you can improve customer service by being reachable to customers at any time, in many languages, and by solving problems. And you reduce costs. According to an IBM study, companies worldwide spend more than $1.3 trillion to handle 256 billion calls from their customers each year. In other words, each service call costs an average of $30. That is huge! Companies are looking for solutions to deliver the same quality in customer service, only at a lower price. Affordable service is also in the interests of customers. But a chatbot that deflects customers instead of helping them doesn’t do anything better – on the contrary. Many companies are therefore looking for new, more effective solutions. Not only in marketing and sales, but also in service. They want and need to digitize and automate more with meaningful self-service solutions.

What should such a solution look like?

Glock: As customers, there are four hurdles that customer service has to overcome. First, we expect problems to be solved quickly, second competently, third at any time, and fourth possible on all channels. Many simple chatbots are based on rudimentary software that can only provide ready-made answers based on precisely asked questions. FAQs or rules with keywords are stored. They do not provide real added value to information on the website or in the app.

A smart solution is only available through the combination of AI-based speech comprehension and the connection of bots to back-end systems. The smart, connected bot can then access an enormous knowledge of the customer’s products, services or personal information from the customer database in real time and is not limited to a question-and-answer catalog.

The smart bot or virtual assistant also has writing access in the backend, i.e. it can change an address or order quantity, create an appointment, cancel a booking, place a callback, etc. Our platform, Cognigy.AI, is already widely used for this purpose. You could also say that earlier bots could provide some general information, later more individual information, and today the bot can self-sufficiently initiate complex business processes. Just like a real employee in the call or contact center.

How much can bots automate for customer service? Will they eventually replace people in contact centers?

Glock: There will always be a need for people to manage the communication and processes in customer service. Bots are good helpers and they are becoming more and more capable. This greatly increases the effectiveness of customer service and also the customer experience for the user. For one thing, the response and business logic behind the bots has to be dictated, hence the rise of the Conversational Designer. And they are always needed for tricky cases or problems where decisions have to be made. Take complaints, for example: a customer is dissatisfied and wants to express his displeasure. As a rule, a person is better suited, as he can, for example, reduce prices or ship new goods out of goodwill. These are individual decisions that a bot cannot make. However, the bot can accept, process and pass on concerns to the right agent in the contact center outside of business hours, who will take care of them the next morning. Customers then have a good feeling, and the agent can work more effectively.

The support of important B2B customers should also be possible face to face with a personal contact person. However, if the business customer simply wants to change an order, query a delivery date or receive a specific product info, the virtual agent can also help – immediately and around the clock.

Keyword AI: What is technologically behind a smart bot? 

Glock: A smart bot can remember things and conduct dialogues like a human being. The technology behind it is called Conversational AI. At its heart is Natural Language Understanding (NLU), i.e. AI-based “understanding” of human language. Understanding means that inputs do not have to be exact but are recognized in a wide range. An example is an address change: A customer could formulate this differently: “I moved to San Francisco”, “change address”, “I have a new address”. There are dozens of ways to say the same thing. AI can assign all utterances to a customer’s intention without all conceivable formulations being programmed or stored. In the case of the statement Move + San Francisco, the bot immediately “remembers” the city as input. For this, the bot is trained in advance.

Once such foundations have been laid, AI can be quickly scaled to other languages and markets. This is immensely important for our mostly international customers. But AI alone is not enough for this. You also need to be able to easily and clearly manage and customize the bot responses in any language, design structured dialogs, connect back-end systems, and implement bots in channels. This has been complicated up to now and can only be implemented with a lot of IT know-how and programming work. With low-code platforms like ours, employees today are able to build their own virtual service agents, continuously improve them, and scale into other markets and channels. This is an elementary step that has only been made possible by technological leaps of the last two years.

Are there specific industries or companies that can use bots particularly well in customer service?

Glock: Every company with a large volume of dialogue across telecommunications, logistics, travel, banks, insurance companies, healthcare and government – all industries are recognizing the added value. There are dozens of useful applications. For example, let’s say I want to submit an insurance claim report on Sunday at midnight or change or query something in the car on the way to the airport.

There are also cases where providers suddenly have to deal with a particularly large volume of dialogue, peak times when many people need urgent information. For example, in the case of home insurance, this happens when there is severe weather. Currently, many people have questions about the COVID-19 vaccination. Handling these peak times can be easily and efficiently cushioned with a Conversational AI platform. A great example is BioNTech, which has implemented virtual agents in multiple languages to meet the current flood of requests from interested parties, customers, doctors, pharmacists, media representatives and suppliers through all channels.

Or take institutions such as the Employment Agency, which engages thousands of people in telephone dialogue every day – on very well-workable topics such as requests, changes and general information needs. Here, bots could greatly increase the quality of customer service. Customers would no longer have to wait long in queues or wait for the clerk to call back, but could resolve their concerns directly with the bot.

How do virtual agents and human agents work hand in hand?

Glock: Often the bot is upstream in order to pass on the right agent directly. Or it acts completely in the background: it listens to a conversation and provides the person in customer service, for example, data on customer history or details about the product specification that is currently being discussed. It makes people smarter and improves their response quality. The customer receives more qualified answers faster. Such so-called agent-assist solutions are very much in demand.

A better-informed service representative is also valuable for up- and cross-selling. He/she can inform the customer of suitable offers in conversation, if reasonable and appropriate. Forrester’s analysts estimate that brands spent about $8 billion more on customer service employees last year than in the previous year, because customer service has to make increasingly qualified statements about increasingly complex products. Assist solutions will therefore continue to boom, as they also help less deeply trained and thus cheaper employees to gain more knowledge.

Are virtual assistants also in demand internally for corporations?

Glock: Clearly. There are also large requests volumes internally. We work with many corporate HR department customers who optimize their internal contact center for employees with our bots. Employees have a lot of questions, currently around Coronavirus, benefits and working from home, but also about processes and regulations. This can all be illustrated in a spoken dialogue via a virtual contact person, also via voice bot.

Are interactions with bots via spoken language the future?

Glock: Definitely! The technology has made a huge leap in voice over the last two years. Language interaction is very easy and natural for us. It’s much easier than filling out form fields on your smartphone and the like. You have your hands-free, you don’t have to look at it or click anywhere. It is a useful form of interaction for many business cases. But the context has to fit.

Such customer service bots must not be lumped together with assistants such as Alexa, Siri or Google Home. They make voice an operating interface, acceptable in broad sections of the population. But the applications are different. An airline’s chatbot can answer everything about air travel, rebook tickets and accept food preferences. Does he need to know the weather at the resort? No! You wouldn’t expect that from a contact center employee.

One last tip?

Glock: Digitization and automation are also changing customer service decisively. Companies, especially those with a large volume of dialogue, should now address the innovative solutions on the market and gain experience in the use of smart chat and voice bots. With a scalable, intuitive platform as a basis, they can start with manageable projects, build up internal expertise and then extend the solution globally in dozens of languages, to other applications and additional channels. Today there are excellent platforms and you need to determine which one is the best fit for your organization.

Thank you for the conversation!

About Cognigy

Cognigy is a global leader in Conversational AI to support customer service automation. Its low-code platform, Cognigy.AI, enables enterprises to automate contact centers for customer and employee communications using intelligent voice- and chatbots. With precise, reliable intent recognition, human-like dialogs and seamless integration into backend systems, Cognigy.AI creates superior user experiences and helps companies reduce support costs. Cognigy.AI is available in SaaS and on-premise environments and supports conversations in any language and on any channel including phone, webchat, SMS and mobile apps. Cognigy’s worldwide client portfolio includes Daimler, Bosch, Henkel, Lufthansa, Salzburg AG and many more. Learn more at

Leave a Comment