AI is a core competence of Consileon. In addition to the pure AI knowledge, we also cover the IT and technical know-how to successfully anchor AI in companies:
Artificial intelligence is suitable for solving abstract numerical examples, as well as for business processes with high numbers of cases where individual consideration of the cases is necessary. One such process is customer communication.
Reinforcement learning – AI in customer relationship management.
In reinforcement learning, a subfield of machine learning, the cognitive agent is given a repertoire of actions with which it can change the state of its environment. In practice, such a state change would be, for example, making a purchase or becoming a more satisfied customer. The agent selects an action according to the situation, receives feedback on the effect of the action, and adjusts its strategy if necessary.
Example: A robot (agent) is supposed to bake you a cake (action). This changes at least two states: the stock of ingredients in the kitchen and your mood depending on how you like the cake. If the taste leaves something to be desired, the robot should know this so that it can make the next cake better. You need to operationalize your feedback (reward) in advance. For example, the robot could measure how much of each piece of cake is left on the plate. Using the information about ingredients and taste judgment, the agent can get closer and closer to your preferences through trial and error. Depending on the number and gradation of actions the robot is allowed to choose from, as well as the degree of freedom you allow it to make choices, this path can be long. In order to reach the goal quickly, experts should preconfigure the agent and fit it into its professional environment.
Used in customer service, the cognitive agent gets to know the preferences of its human counterpart better with each interaction. By adapting, it optimizes the rules of address. Ideally, the agent, and thus your company, becomes a trustworthy partner for the customer. People know and appreciate each other. Almost like in the corner store of the 19th century, the company treats the customer as an individual again.
In the image above, we have applied this scheme to relationship marketing (CRM). The state to be changed in the environment of the agent (robot) is a customer profile. Sensors are instruments that can be used to detect changes in the state; in our case, the classic performance measurement is used for this purpose. Finally, actuators (actuating elements or “switches”) are the means of addressing the customer on all interaction channels. The cognitive agent replaces or extends the old CRM system. To enable the system to optimize the interaction with the customer according to his “taste”, a feedback system is built in. The feedback is based on the business goals. Such a system is more efficient and can be linked more closely and flexibly to goals than classic commission models, which attempt a comparable optimization with sales personnel as human “agents”.
Many companies have long been using sensors such as success measurement or actuators such as a differentiated customer approach. In the course of digitization, cognitive agents are becoming more and more efficient because the precision of sensors and actuators is increasing, as is their diversity. Human “agents,” on the other hand, will hardly be able to collect and utilize all the information.
Did we spark your interest?
We are familiar with the use of cognitive agents for CRM purposes. In a preliminary discussion, we clarify which approach suits your company. In workshops, we sensitize managers and specialists to identify potential AI fields of application in their own companies. We help you build up internal AI expertise and solve specific use cases. We look forward to accompanying you on this journey from the very beginning.
