16. December 2020 By Jan Heuker
Artificial Intelligence leads to real service
Artificial intelligence within service processes will be used in 70 per cent of customer interactions by 2022 (according to Gartner's Magic Quadrant for the CRM Customer Engagement Centre, May 2018). That is a considerable improvement on the 15 percent in 2018. And that's much needed, because everyone is horrified by helplines with long waiting times, continually entering the same data and ultimately being left with a mess.
That the service of helplines can also be different can be read in this article about the bots Maya and Jim, who close policies and handle claims at insurer Lemonade without human intervention. But there are of course many other examples. This AI research also shows that artificial intelligence in service processes is a real customer booster. How that works exactly, we will explain here.
Unburdening service teams
AI applications are capable of taking over individual aspects of customer communication, thus relieving the burden on service teams. They can also improve the performance of individual employees and control the overall process of service communication. In this way, AI applications of this kind create more transparency for all those involved in the process.
The ability of AI applications to process text and speech plays a central role in all this. Think of technology like Natural Language Processing (NLP), with which the system recognises the connections and meanings of spoken words. Or Natural Language Generation (NLG), which uses machine learning to apply stylistic and accentuating features to automatically generated text.
What makes chatbots with artificial intelligence so special is that they enable customers to communicate with them in natural language. Well-trained chatbots are suitable for processing enquiries, for example: they explain the differences between product variants, provide information about the delivery status of an order or offer help in planning a trip to an event. Advanced chat applications are capable of highly intelligent dialogues.
But chatbots are not only useful at the front end. They also provide direct support to service staff in their work. For example, as trainers and sparring partners for new employees or for learning new subjects. Chatbots then take on the role of questioners in training sessions. The evaluation and assessment of the dialogues can then also be carried out automatically.
Chatbots can also support the daily work of service staff, ultimately resulting in better advice. To this end, chatbots continuously analyse the ongoing dialogue. On the basis of the recognised content, they provide the employee with appropriate information from contract documents, chat protocols, orders, manuals, leaflets, instructions or frequently asked questions (FAQs). This information is pre-filtered, summarised and prioritised.
Service agents can use this information directly during the call, and in this way they can provide faster and more expert advice. And all this without the employee having to search through all the available information sources himself.
Nowadays, customers contact an organisation via multiple channels, such as telephone, e-mail, contact forms, letters, Twitter or Facebook. Each channel poses new challenges for those responsible. Even the recording and forwarding of questions is still often a manual task. For example, individual employees decide how to enter a particular query into which system (CRM, ticket system or company social network). This task is crucial for the quality of the entire service process and requires some understanding of the organisation, yet it is often performed by insufficiently prepared and inexperienced employees. This shortcoming, however, can be overcome with AI.
Another obstacle on the road to better service processes is compartmentalised thinking: sales, marketing, accounting, logistics. Each system contains only part of a customer's data. What is lacking is an integrated overview of the entire process and all service procedures. AI technologies are capable of making all data into a whole. Here too, AI technologies show their power in processing texts and language.
The overall process that an AI-based chatbot goes through is that it first retrieves all incoming requests or messages from all the individual input channels. The application then identifies other queries from the same person or customer and adds this information to the metadata. At the same time, the content is interpreted into unstructured information, such as e-mail texts or social media articles.
Based on the metadata, the AI application sends the incoming message to the appropriate processing system. The system tracks processing progress throughout the process; it recognises new requests on the same topic or by the same person and automatically assigns them.
The AI solution also monitors all communication about a process and records employee questions and responses to a request. The system thus provides insight not only into the individual process, but into the entire workflow. On request, it also determines the central key figures, such as the average processing time, the number of questions on individual topics or the distribution of questions over time. These processes prevent unnecessary loops in the process.
By integrating artificial intelligence into the service process, you always have all the necessary data at hand, when it is needed, and everyone involved can therefore provide the right information. Artificial intelligence makes service delivery not only more efficient but also more human. An AI-driven system is able to understand requests better and can generate meaningful answers faster. And this gives the service worker more time to focus on the person asking for help. After all, it's not about technology, but about people and their needs.