AI, Artificial Intelligence, or machine learning. All hot topics that appear in many different ways in marketing, but what exactly is it all about?
In summary, it is about the fact that very old statistical and mathematical models and methods are being used to group and analyze information or predict the future. What has changed is the ability to do this in real time at great speed and process more and more information on a much larger scale. , This gives us many more use scenarios with greater background information.
Another current buzz word is robotics, often referring to software, and no longer just physical machines. Basically, software robotics is traditional IT thinking, automating manual recurring processes and so releasing people's working time for other tasks.
Let’s cut the topic into a few bite sized pieces and digest these further.
Let’s check what can be done prior to the customer’s contact with customer service. In the build-up phase of artificial intelligence models, a bunch of information sources such as Business Intelligence-data, CRM-data and other predictions about how to improve customer service can be used. For example, history and other background information can be used to prioritize outbound calling lists or ticket order handling and enhance employee shift planning.
It is possible to predict who should be called at certain times, how large quantities of incoming contact are to be expected for certain time periods and how much resource is needed during the day for different services. The difference to the old statistical analysis is that the models are not constructed and utilized one-off and manually, rather machine learning can update the models in real-time.
When a customer contact comes in, adaptive models can enhance smart routing by changing routing rules in real time. In addition to identifying customer relationships and utilizing background information to guide calls to the right person, many other kinds of information particularly those which update fast can be used. This information is utilized by artificial intelligence to guide identified customers in the right direction. For example in a situation where a customer order is at a certain phase, the calendar takes account of holiday seasons or if it starts to freeze outside.
Before contact from customers is transferred to a customer service representative, either when waiting in a call queue, straight away when the email has arrived or when the chat is started, robotics and artificial intelligence will rush to help.
They can ensure that customer receives real-time information about the a waiting time in call queue and is even able to talk with a software robot equipped with artificial intelligence.
The robot is able to map out the situation and present solution options to the caller. If the customer receives the necessary information or gets the process started before the customer service representative is free to answer, time is saved and the issue is handled in shorter time.
Whilst serving the customer, software robots and artificial intelligence can provide background information, such as reports, information from CRM and from other external systems to help in the customer service situation. Artificial intelligence can also propose solutions to the customer’s questions, identify and propose additional sales opportunities, and change a predefined walkthrough guide or “script" of the call.
Enhanced processing of the data has also brought opportunities for real-time voice recognition. During a call, it’s possible to identify (even more objectively than the customer service representative him/herself), the irritability in a customer's voice or other types of emotions. In this case, the call can be redirected in a positive direction or even escalated to the customer service representative’s supervisor.
After the contact, classification and grouping can be automated and analyzed based on the content of communication. Also, after the contact, robotics and artificial intelligence can process the automated part of post-work, such as bidding management and entering data into other systems.
How then to start building these artificial intelligence and robotics benefits to help customer service? It's a good idea to start with simple models and at those points where the highest cost-benefit ratio is achieved.
For example, if the problem is with the quantity of incoming emails, and that in most cases is very repetitive work, email bots can be a good pilot program. If, on the other hand, the outbound call campaigns are conducted with poor success, analytics based on the outbound call lists can help. Intelligent routing may, in turn, help to release company resources to serve customers exactly in their own area of expertise and this way reduce bouncing the customer to different places.
Whatever the need, we are happy to discuss more with you about how to implement intelligent solutions with our partners to support your customer service success.
Writer: Mikko Muurinen, Chief Analytics Officer Benemen Oy | Linkedin profile
Based in Helsinki, Finland Mikko is responsible for Business Intelligence solutions in Benemen. Since the turn of the millennium, Mikko has been working with Business Intelligence, both in the public and private sectors. Mikko has experience in IT, sales, product management and management positions. The statistical-mathematical background of his education explains Mikko's passion for all the data, but it is still unclear why he so loves chocolate and the colour orange so much.